Abstract
Introduction
Hip fracture in geriatric patients has a substantial economic impact and represents a major cause of morbidity and mortality in this population. At our institution, a regional anesthesia program was instituted for patients undergoing surgery for hip fracture. This retrospective cohort review examines the effects of regional anesthesia (from mainly after July 2007) versus general anesthesia (mainly prior to July 2007) on morbidity, mortality and hospitalization costs.
Methods
This retrospective cohort study involved data collection from electronic and paper charts of 308 patients who underwent surgery for hip fracture from September 2006 to December 2008. Data on postoperative morbidity, in-patient mortality, and cost of hospitalization (as estimated from data on hospital charges) were collected and analyzed. Seventy-three patients received regional anesthesia and 235 patients received general anesthesia. During July 2007, approximately halfway through the study period, a regional anesthesia and analgesia program was introduced.
Results
The average cost of hospitalization in patients who receive surgery for hip fracture was no different between patients who receive regional or general anesthesia ($16,789 + 631 v. $16,815 + 643, respectively, p = 0.9557). Delay in surgery and intensive care unit admission resulted in significantly higher hospitalization costs. Age, male gender, African-American race and intensive care unit admission were associated with increased in-hospital mortality. In-hospital mortality and rates of readmission are not statistically different between the two anesthesia groups.
Conclusions
There is no difference in postoperative morbidity, rates of re-hospitalization, in-patient mortality or hospitalization costs in geriatric patients undergoing regional or general anesthesia for repair of hip fracture. Delay in surgery beyond 3 days and intensive care unit admission both increase cost of hospitalization.
Introduction
Hip fracture is a common, potentially devastating event for geriatric patients with approximately 1.6 million hip fractures occurring annually worldwide (1) and a projected 6.2 million hip fractures annually by the year 2050 (1). In 2004 in the US alone, an estimated 387,000 patients older than 65 were hospitalized due to falls and about 37% of these patients sustained hip fractures (2). The mean cost of hospitalization for patients with hip fractures was $18,638 with an average length of hospital stay of 7 days (2).Hip fracture has consistently been found to be associated with a high morbidity and mortality (3-5) with reported mortality remaining stable in the last 40 years (6). A higher mortality has been associated with increasing age and male gender (7,8).
At the University of Florida and Shands Hospital (UF&Shands), we have performed an increasing number of surgeries for geriatric patients with fragility hip fractures since the opening of our level I trauma center in 2005. Prior to July 2007, most surgeries for hip fractures were performed using general anesthesia. A shift in the involvement of the Acute Pain Medicine Service in July 2007 resulted in a change in intraoperative anesthetic management of geriatric, debilitated patients, which involved placement of either single injection spinals, continuous spinal catheters, or epidural catheters for intra-operative anesthesia and femoral nerve blocks or lumbar paravertebral blocks for postoperative analgesia. We also introduced a specialized team of medical personnel dedicated to postoperative analgesia. Subsequently we conducted a retrospective cohort study to examine patient characteristics, cost of hospitalization (estimated from charges of hospitalization), readmission rates and in-hospital mortality for these patients. Our hypothesis was that the use of regional anesthesia would reduce pulmonary morbidity, and therefore reduce the need for critical care intervention postoperatively, length of hospital stay, and cost of hospitalization.
Methods
Data Source
Investigators received institutional review board approval and a waiver of informed consent prior to gathering and analyzing data. Patients who underwent surgery for hip fractures (percutaneous pinning, open reduction internal fixation or bipolar arthroplasty) from September 2006 to December 2008 were identified using Current Procedure Terminology (CPT) codes. We arbitrarily chose this period of time for the equal number of months before and after institution of regional anesthesia for hip fracture repair, and to allow for up to at least one year follow-up after the last discharge date to assess one-year mortality. The hospital’s Decision Support Services Analyst provided the authors with demographic data, charges for surgery and hospitalization, hospital characteristics, outcomes (eg, in-hospital mortality, length of stay, number of ICU days, ICU charge, discharge facility), and insurance status.
Hospital charge data included physician fees, procedural charges, and equipment charges. Cost of hospitalization was estimated by applying the ratio of cost-to-charge for urban hospitals in the South Atlantic division (Healthcare Cost and Utilization Project, 2008 National Statistics) (9) to the amount charged for hospitalization. International Classification of Diseases, 9th Edition (ICD-9) codes for up to 15 diagnoses and procedures are listed for each admission. We reviewed the perioperative anesthesia records and recorded type of intraoperative anesthesia (general versus regional) and postoperative analgesia (systemic opioids versus peripheral or neuraxial continuous nerve block). General anesthesia included the use of inhalational with or without intravenous agents for induction and maintenance of anesthesia. Regional anesthesia included the use of a neuraxial technique (single injection spinal, continuous spinal, or continuous epidural anesthesia) with or without the use of a continuous peripheral nerve block. Of note, patients with failed neuraxial blockade who subsequently underwent surgery using general anesthesia were included in the GA group for analysis.
Patient Population
All surgeries were performed at a single hospital in an academic setting. Patients were included in the retrospective review if their age was greater than or equal to 65 and if they sustained a low-impact hip fracture requiring surgical stabilization.
We identified 319 patients who underwent a total of 324 surgeries for fragility hip fracture at UF&Shands between September 2006 and December 2008. Patients who subsequently had a second low-impact hip fracture requiring stabilization during the time period of the study who may or may not have the same type of intraoperative anesthetic (n=5) were not included in the analysis. Patients with preexistent end-stage renal disease requiring dialysis (ESRD) (n=6) were not included in the analysis. Of the 308 patients included in the statistical analysis, 73 received regional anesthesia and 235 received general anesthesia intraoperatively.
Outcomes
We reviewed admission and discharge summaries from the hospital’s electronic records to acquire information about hospital readmission within thirty days of discharge and reason for readmission. Information regarding number of days of hospitalization, days from admission to surgical stabilization, ICU length of stay, days on mechanical ventilation, in-patient mortality, International Classification of Disease, Ninth revision (ICD-9) codes for admission diagnoses, condition and disposition at discharge, and discharge facility were provided by the hospital’s Decision Support Services Analyst.
Statistical Analysis
The primary outcomes of interest in this study were in-hospital mortality, in-hospital complications, hospital length of stay, and total costs (as estimated from hospital charges). All categorical outcomes were adjusted using multiple logistic regression models that included age, race, gender, anesthesia type, day of admission (weekend vs. weekday), admitting service, surgery delay, length of stay, ICU admission, presence of cardiovascular disease, chronic obstructive pulmonary disease, cancer, diabetes, and dementia. For continuous outcomes we used linear regression models adjusted for the same factors as above. In addition, because this was an observational study, to adjust for the lack of randomization that could result in large differences on the observed and unobserved patient characteristics between the general anesthesia and regional anesthesia groups, we used a propensity scoring method to create an analysis that resembled what would occur if the anesthesia had been randomly assigned. This method allowed us to reduce the bias that would result due to the lack of randomization of subjects into anesthesia groups to properly estimate the effect of anesthesia type on length of stay and hospital charges. A propensity score, which is the probability of the patient being assigned to general anesthesia rather than regional anesthesia conditional on the patient’s demographic characteristics and comorbidities, was calculated for each patient using the logistic regression. The propensity score weights, which are the inverse of the propensity scores, were calculated and used in the analyses as weight (10). Propensity score-weighted linear regression models were fitted to compare the effect of anesthesia type on the cost of hospitalization and length of stay. Least squares means were compared for the propensity score adjusted and the unadjusted multiple regression models. Linear regressions were performed using untransformed costs because mean estimates obtained were similar to those of alternatives approaches that better account for skewed data. The coefficient of determination, R2, was calculated as an indicator of the proportion of variability explained by each model. Length of stay and hospital charges were compared among subgroups with analysis of variance and Kruskal-Wallis test, then modeled in linear regressions. For univariate analysis of categorical variables we used Pearson χ2 analyses or Fisher’s exact test as appropriate. The goodness of fit of the logistic regression models was assessed with the Hosmer–Lemeshow test, and c-statistics evaluated the discriminatory capability of the models. All significance tests were 2-sided, with an α of .05, which we considered to be statistically significant. Statistical analyses were performed with SAS (v.9.2, Cary, NC).
Results
We enrolled a total of 323 patients who underwent hip fracture surgery for fragility hip fracture of whom 308 were eventually submitted to the analysis. Univariate analysis comparing patient characteristics between the GA and RA group are presented in Table 1. Proportions of in-hospital mortality, nursing home discharge, and re-hospitalization and median length of stay, and estimated cost of hospitalization among subgroups that are determined by baseline demographics and clinical characteristics are presented in Table 2. Of the 308 patients, 73 received regional anesthesia and 235 received general anesthesia. Frequency and type of regional anesthesia and analgesia as well as surgical procedure is presented in Tables 3 and 4. The majority of RA patients (96 %) received continuous spinal anesthesia. There were predominantly more females (n=229) than males (n=79), and 188 patients were older than 80 years of age.
Table 1.
Comparison of demographics and comorbidities of patients in general and regional anesthesia groups
| No . (%) of Patients |
||||
|---|---|---|---|---|
| Overall | General Anesthesia (n=235) |
Regional Anesthesia (n=73) |
p-value | |
| Overall | 308 | |||
|
| ||||
| Age (years) ( >=80) | 186 | 132 (56%) | 54 (74%) | 0.0066 |
| Male Gender | 79 | 64 (27%) | 15 (21%) | 0.2532 |
| African-American race | 22 | 18 (8%) | 4 (5%) | 0.5275 |
| Weekend admission | 83 | 74 (31%) | 9 (12%) | 0.0013 |
| Medicine admitting service | 88 | 68 (29%) | 20 (27%) | 0.7993 |
| Cardiovascular Disease | 151 | 119 (51%) | 32 (44%) | 0.3098 |
| Diabetes | 66 | 57 (24%) | 9 (12%) | 0.0301 |
| Chronic Obstructive Pulmonary Disease | 85 | 66 (28%) | 19 (26%) | 0.7312 |
| Cancer | 20 | 13 (6%) | 7 (10%) | 0.2742 |
| Dementia | 36 | 30 (13%) | 6 (8%) | 0.2909 |
| Chronic renal disease | 53 | 40 (17%) | 13 (18%) | 0.8763 |
| Postoperative Intensive Care Unit | 64 | 53 (23%) | 11 (15%) | 0.1686 |
| Postoperative Mechanical Ventilation | 18 | 17 (6%) | 1 (1%) | 0.0841 |
| Postoperative Regional Analgesia | 128 | 64 (27%) | 64 (88%) | <0.0001 |
| Surgery Day Delay (days) | 0.2748 | |||
| 0-1 | 236 | 175 (75%) | 61 (85%) | |
| 2-3 | 55 | 46 (20%) | 9 (12%) | |
| >3 | 17 | 14 (6%) | 3 (4%) | |
Table 2.
Measures of Health Care Use Related to Hip Stabilization Surgery, Patients Aged 65 or Older
| No . (%) of Patients |
Median (Q1, Q3) |
|||||
|---|---|---|---|---|---|---|
| Overall | Inhospital Mortality |
Nursing Home Discharges(n=297) |
Rehospitalization (n=297) |
Length of Stay (days) |
Hospital Cost ($) | |
|
|
|
|||||
| Overall | 308 | 11 (3.57) |
173
(58.25) |
59 (19.87) | 5 (4, 7) |
14689 (11823,
18697) |
|
| ||||||
| Age (years) | ||||||
| 65-79 | 122 | 2 (1.64) | 55 (45.83)* |
23 (19.17) | 6 (4,8) | 15163 (12378, 20095) |
| >=80 | 186 | 9 (4.84) | 118 (66.67) | 36 (20.34) | 5 (4,7) | 14161 (11677, 17923) |
|
| ||||||
| Gender | ||||||
| Female | 229 | 4 (1.75)* | 126 (56.00) |
44 (19.56) | 5 (4,7) | 13958 ( 11731, 17764)* |
| Male | 79 | 7 (8.86) | 47 (65.28) | 15 (20.83) | 6 (4,8) | 16039 (12948, 21133) |
|
| ||||||
| Race | ||||||
| African-American | 22 | 3 (13.64)* | 11 (57.89) | 2 (10.53) | 6 (4,8) | 16780 (13571, 20067) |
| Other | 286 | 8 (2.80) | 162 (58.27) |
57 (20.50) | 5 (4,7) | 14601 (11731, 18642) |
|
| ||||||
| Anesthesia type | ||||||
| General anesthesia | 235 | 9 (3.83) | 125 (55.31) |
42 (18.58) | 5 (4,7) | 14724 ( 11731, 18866) |
| Regional anesthesia | 73 | 2 (2.74) | 48 (67.61) | 17 (23.94) | 5 (4,7) | 13812 (12151, 18480) |
|
| ||||||
| Admission Day of Week | ||||||
| Weekend | 83 | 4 (4.82) | 40 (50.63) | 19 (24.05) | 5 (4,8) | 14333 (10884, 19511) |
| Weekday | 225 | 7 (3.11) | 133 (61.01) |
40 (18.35) | 6 (4,7) | 14718 (12246, 18646) |
|
| ||||||
| Admitting Service Type | ||||||
| Medicine | 88 | 7 (7.95)* | 51 (62.96) | 21 (25.93) | 7 (5, 10)* | 17017 (14106, 21492)* |
| Surgery | 220 | 4 (1.82) | 122 (56.48) |
38 (17.59) | 5 (4, 6.5) | 13681 (11463, 16551) |
|
| ||||||
| Cardiovascular Disease | ||||||
| Yes | 151 | 8 (5.30) | 91 (63.64) | 32 (22.38) | 6 (4, 9)* | 16039 (13299, 21133)* |
| No | 157 | 3 (1.91) | 82 (53.25) | 27 (17.53) | 5 (4, 7) | 13085 (10921, 16312) |
|
| ||||||
| Diabetes | ||||||
| Yes | 66 | 0 (0) | 41 (62.12) | 20 (30.30)* | 6 (4, 10)* |
15371 (12929, 21331) |
| No | 242 | 11 (4.55) | 132 (57.14) |
39 (16.88) | 5 (4,7) | 14537 (11710, 18365) |
|
| ||||||
| COPD | ||||||
| Yes | 85 | 3 (3.53) | 50 (60.98) | 23 (28.05)* | 5 (4,7) | 14902 (12286, 19174) |
| No | 223 | 8 (3.59) | 123 (57.21) |
36 (16.74) | 5 (4,8) | 14623 (11677, 18646) |
|
| ||||||
| Cancer | ||||||
| Yes | 20 | 0 (0) | 10 (50.00) | 1 (5.00) | 7.5 (5, 14)* |
19923 (14127, 24423)* |
| No | 288 | 11 (3.82) | 163 (58.84) |
58 (20.94) | 5 (4, 7) | 14459 (11766, 18082) |
|
| ||||||
| Dementia | ||||||
| Yes | 36 | 1 (2.78) | 26 (74.29) | 9 (25.71) | 5 (4, 7) | 15119 (11776, 18973) |
| No | 272 | 10 (3.68) | 147 (56.11) | 50 (19.08) | 5 (4, 7) | 14647 (11823, 18697) |
|
| ||||||
| CKD | ||||||
| Yes | 53 | 2 (3.77) | 32 (62.75) | 13 (25.49) | 6 (4,9)* | 15306 (12659, 20646) |
| No | 255 | 9 (3.53) | 141 (57.32) |
46 (18.70) | 5 (4, 7) | 14553 (11710, 18286) |
|
| ||||||
| ICU | ||||||
| Yes | 64 | 8 (12.5)* | 35 (62.50) | 17 (30.36)* | 9 (6, 13.5)* |
22025 (17539, 30807)* |
| No | 244 | 3 (1.23) | 138 (57.26) |
42 (17.43) | 5 (4, 7) | 13594 (11398, 16375) |
|
| ||||||
| Surgery Day Delay (days) | ||||||
| 0-1 | 236 | 5 (2.12)* | 135 (58.44) |
43 (18.61) | 5 (4, 7)* | 13767 (11398, 16509)* |
| 2-3 | 55 | 3 (5.45) | 31 (59.62) | 52 (25.00) | 7 (6, 10) | 17287 (13795, 21902) |
| >3 | 17 | 3 (17.65) | 7 (50.00) | 3 (21.43) | 12 (9,13) | 22868 (20095, 25488) |
Significant differences among subgroups, p-value<0.05
Table 3.
Frequency of Anesthesia
| Frequency (%) | |
|---|---|
| Type of Anesthesia | |
| General anesthesia | 222 (72%) |
| Failed regional anesthesia, conversion to general anesthesia | 13 (4%) |
| Continuous spinal catheter | 70 (23%) |
| Local with sedation | 1 (0.3%) |
| Single injection spinal | 1 (0.3%) |
| Combined spinal/epidural | 1 (0.3%) |
Table 4.
Propensity Score-Weighted Linear Regression Models on Hospital Cost and Length of Stay
| R2=75.4% | R2=33.18% | |||
|---|---|---|---|---|
| Hospital Cost ($) | Length of Stay (days) | |||
|
|
|
|||
| Predictors |
Estimate
(SE) |
p-value |
Estimate
(SE) |
p-value |
| Age (years) | 65 (52) | 0.2140 | 0.05 (0.05) | 0.3190 |
| Male Gender (vs. Female) | 460 (539) | 0.3947 | 0.93 (0.49) | 0.0562 |
| African-American Race( vs. Other) | 28 (925) | 0.9759 | −1.10 (0.83) | 0.1860 |
| General Anesthesia (vs. Regional) | 25 (457) | 0.9557 | 0.19 (0.41) | 0.6438 |
| Weekend Admission ( vs. Weekday) | 429 (527) | 0.4159 | −1.27 (0.47) | 0.00 75* |
| Medicine Admitting Service ( vs. Surgery) | −1128 (549) | 0.040 7* |
0.36 (0.56) | 0.5205 |
| Cardiovascular Disease (Yes vs. No) | 410 (504) | 0.4161 | 0.75 (0.45) | 0.0974 |
| COPD (Yes vs. No) | 921 (521) | 0.0850 | −0.72 (0.47) | 0.1231 |
| Dementia (Yes vs. No) | 645 (708) | 0.3630 | −0.88 (0.64) | 0.1710 |
| Cancer (Yes vs. No) | 2290 (1086) | 0.035 8* |
2.46 (0.97) | 0.01 16* |
| Diabetes (Yes vs. No) | −563 (582) | 0.3339 | 0.85 (0.52) | 0.1036 |
| ICU (Yes vs. No) | 4389 (655) | <.000 1* |
2.95 (0.56) | <.00 01* |
| Inhospital Mortality (Yes vs. No) | 2873 (1368) | 0.036 5* |
−0.32 (1.23) | 0.7947 |
| Surgery Day Delay (hours) (2-3 vs. 0-1) | NA | NA | 6.87 (1.04) | <.00 01* |
| (>3 vs. 0-1) | NA | NA | 1.52 (0.63) | 0.01 68* |
| Length of Stay (day) | 1360 (61) | <.000 1* |
NA | NA |
Significant differences among subgroups, p-value<0.05
The odds ratios (OR) for in-hospital mortality, nursing home discharge, and re-hospitalization rates obtained by multiple logistic regression (c-statistic=0.949) are reported in Table 5. A higher in-hospital mortality was associated with increasing age (OR, 1.6), male gender (OR, 8.1), African-American race (OR, 16.1) and ICU admission (OR, 24.2). Anesthesia type was not found to be significantly associated with in-hospital mortality. The OR of nursing home discharge was significantly higher for patients with dementia (OR, 2.6), whereas the OR of re-hospitalization was significantly higher for patients with COPD (OR, 2.3) or diabetes (2.5). Cardiovascular disease was associated with a higher overall complication rate and higher cardiovascular and pulmonary complication rate. Delay in surgery beyond 3 days was associated with higher overall complication rate and a higher cardiovascular complication rate (Table 5).
Table 5.
Regression on overall complications, pulmonary complications, and cardio complications
| c-statistic | 0.7290 | 0.7730 | 0.8210 | |||
|---|---|---|---|---|---|---|
| Hosmer-Lemeshow Goodness-of-Fit Test | 0.8592 | 0.4780 | 0.9513 | |||
| Complications |
Pulmonary
Complications |
Cardio Complications | ||||
| Predictors |
OR (CI for
OR) |
p-value |
OR (CI for
OR) |
p-value |
OR (CI for
OR) |
p-value |
| Age (years) | 1.0 (0.9, 1.1) |
0.5822 | 1.0 (0.9, 1.1) |
0.9472 | 1.0 (0.8,1.1) | 0.4485 |
| Male Gender (vs. Female) | 1.8 (0.9, 3.6) |
0.0972 | 2.3 (1.0, 5.2) |
0.0520 | 1.7 (0.6, 4.8) | 0.2799 |
| African-American Race( vs. Other) | 0.7 (0.2, 2.6) |
0.5462 | 1.5 (0.4, 6.2) |
0.5752 | 0.4 (0.0,3.9) | 0.4409 |
| General Anesthesia (vs. Regional) | 1.7 (0.7, 4.1) |
0.2056 | 2.2 ( 0.7, 7.2) |
0.1751 | 1.7 (0.4, 6.3) | 0.4534 |
| Weekday Admission ( vs. Weekend) | 2.1 (0.9, 4.8) |
0.0712 | 1.9 (0.7, 5.1) |
0.2287 | 1.8 (0.6, 5.7) | 0.3069 |
| Surgery Admitting Service ( vs.Medicine) | 1.4 (0.7, 3.2) |
0.3730 | 2.0 (0.7, 5.3) |
0.1758 | 2.4 (0.7, 8.0) | 0.1427 |
| Cardiovascular Disease (Yes vs. No) | 3.1 (1.5, 6.4) |
0.0019* | 5.1 (1.9, 13.7) |
0 .001* |
8.1 (2.2, 30.4) | 0.001 9* |
| COPD (Yes vs. No) | 1.5 (0.8, 3.0) |
0.2609 | 1.2 (0.5, 2.8) |
0.7336 | 2.3 (0.9, 6.1) | 0.0883 |
| Dementia (Yes vs. No) | 2.1 (0.8, 5.2) |
0.1165 | 1.9 (0.6, 5.8) |
0.2811 | 1.9 (0.5, 7.0) | 0.3465 |
| Cancer (Yes vs. No) | 1.1 (0.3, 3.9) |
0.9153 | 2.4 (0.6, 9.6) |
0.2152 | 1.5 (0.3, 8.1) | 0.6577 |
| Diabetes (Yes vs. No) | 0.8 (0.3, 1.7) |
0.5211 | 0.7 (0.3, 1.9) |
0.5180 | 1.1 (0.4, 3.3) | 0.8728 |
| Surgery Day Delay (hours) (2-3 vs. 0-1) | 1.2 ( 0.5, 3.0) |
0.6670 | 1.2 (0.4, 3.4) |
0.8012 | 0.7 (0.2, 3.1) | 0.6801 |
| (>3 vs. 0-1) | 3.6 (1.1, 12.2) |
0.038 7* |
2.4 (0.5, 10.8) |
0.2419 | 6.4 (1.5, 28.5) | 0.014 3* |
Significant differences among subgroups, p-value<0.05
Table 6 shows the propensity-score weighted regression analysis results for estimated hospital cost and length of stay. Although the median estimated cost of hospitalization was higher for patients admitted to the medicine service (as compared to surgical admitting service), admission by the department of internal medicine resulted in a lower cost of hospitalization after adjusting for the demographic characteristics and comorbidities than the admitting service of the department of surgery. Patients with a diagnosis of cancer, patients who require an admission to the intensive care unit (ICU) and patients who died or stayed longer during hospitalization had higher estimated costs of hospitalization. Mean estimated cost of hospitalization ($16,789 vs $16,815) and mean length of stay (6.4 vs 6.6 days) were not significantly different between the patients undergoing hip fracture surgery under regional anesthesia and general anesthesia (Table 7 and 8). Table 9 summarizes the effect of anesthesia type on complications and health care use and there are no significant differences between the two anesthesia groups.
Table 6.
Propensity Score-Weighted Linear Regression Models on Hospital Cost and Length of Stay
| R2=75.4% | R2=33.18% | |||
|---|---|---|---|---|
| Hospital Cost ($) | Length of Stay (days) | |||
|
|
|
|||
| Predictors |
Estimate
(SE) |
p-value |
Estimate
(SE) |
p-value |
| Age (years) | 65 (52) | 0.2140 | 0.05 (0.05) | 0.3190 |
| Male Gender (vs. Female) | 460 (539) | 0.3947 | 0.93 (0.49) | 0.0562 |
| African-American Race( vs. Other) | 28 (925) | 0.9759 | −1.10 (0.83) | 0.1860 |
| General Anesthesia (vs. Regional) | 25 (457) | 0.9557 | 0.19 (0.41) | 0.6438 |
| Weekend Admission ( vs. Weekday) | 429 (527) | 0.4159 | −1.27 (0.47) | 0.00 75* |
| Medicine Admitting Service ( vs. Surgery) | −1128 (549) | 0.040 7* |
0.36 (0.56) | 0.5205 |
| Cardiovascular Disease (Yes vs. No) | 410 (504) | 0.4161 | 0.75 (0.45) | 0.0974 |
| COPD (Yes vs. No) | 921 (521) | 0.0850 | −0.72 (0.47) | 0.1231 |
| Dementia (Yes vs. No) | 645 (708) | 0.3630 | −0.88 (0.64) | 0.1710 |
| Cancer (Yes vs. No) | 2290 (1086) | 0.035 8* |
2.46 (0.97) | 0.0116* |
| Diabetes (Yes vs. No) | −563 (582) | 0.3339 | 0.85 (0.52) | 0.1036 |
| ICU (Yes vs. No) | 4389 (655) | <.000 1* |
2.95 (0.56) | <.00 01* |
| Inhospital Mortality (Yes vs. No) | 2873 (1368) | 0.036 5* |
−0.32 (1.23) | 0.7947 |
| Surgery Day Delay (hours) (2-3 vs. 0-1) | NA | NA | 6.87 (1.04) | <.00 01* |
| (>3 vs. 0-1) | NA | NA | 1.52 (0.63) | 0.01 68* |
| Length of Stay (day) | 1360 (61) | <.000 1* |
NA | NA |
Significant differences among subgroups, p-value<0.05
Table 7.
Mean Hospital Cost ($) by Anesthesia Type : Unadjusted vs. Propensity Score Adjusted Models
| Model | Anesthesia Type | Hospital Cost ($) | 95 % Confidence Limits |
p-value for difference |
Difference Between Means |
95% Confidence Limits for the Difference |
||
|---|---|---|---|---|---|---|---|---|
| Propensity Score | General Anesthesia | 16815 | 16172 | 17457 | 0.9557 | 25 | −875 | 925 |
| Adjusted Model | Regional Anesthesia | 16789 | 16158 | 17420 | ||||
|
| ||||||||
| Unadjusted | General Anesthesia | 16816 | 16246 | 17386 | 0.9209 | −61 | −1263 | 1142 |
| Model | Regional Anesthesia | 16877 | 15835 | 17919 | ||||
Table 8.
Mean Length of Stay (days) by Anesthesia Type : Unadjusted vs. Propensity Score Adjusted Models
| Model | Anesthesia Type | Length of Stay (days) |
95 % Confidence Limits |
p-value for difference |
Difference Between Means |
95% Confidence Limits for the Difference |
||
|---|---|---|---|---|---|---|---|---|
| Propensity Score | General Anesthesia | 6.6 | 6.0 | 7.1 | 0.6438 | 0.2 | −0.6 | 1.0 |
| Adjusted Model | Regional Anesthesia | 6.4 | 5.8 | 6.9 | ||||
|
| ||||||||
| Unadjusted | General Anesthesia | 6.7 | 6.2 | 7.1 | 0.7331 | −0.2 | −1.2 | 0.8 |
| Model | Regional Anesthesia | 6.8 | 6.0 | 7.7 | ||||
Table 9.
Complications and Health Care Use as a function of anesthesia type
| Adjusted Mean | 95% Confidence Interval) |
||
|---|---|---|---|
| Hospital Cost, adjusted mean, US $ | Regional Anesthesia | 16789 | (16158, 17420) |
| General Anesthesia | 16815 | (16172, 17457) | |
|
| |||
| Length of Stay, adjusted mean, d | Regional Anesthesia | 6.38 | (5.81, 6.93) |
| General Anesthesia | 6.57 | (5.99, 7.14) | |
|
| |||
| Odds Ratio |
95% Confidence
Interval) |
||
| Complications (Yes/No) | General vs. Regional | 1.7 | (0.7, 4.1) |
| Pulmonary complications (Yes/No) | General vs. Regional | 2.2 | (0.7, 7.2) |
| Cardio complications (Yes/No) | General vs. Regional | 1.7 | (0.4, 6.3) |
|
| |||
| Inhospital Mortality (Yes/No) | Regional vs. General | 1.0 | (0.1, 6.7) |
| Nursing Home Discharges (Yes/No) | Regional vs. General | 1.7 | (0.9, 3.1) |
| Rehospitalization (Yes/No) | Regional vs. General | 2.0 | (1.0, 4.0) |
Discussion
In this study, we found that there is no added cost to the institution for use of regional anesthesia, predominantly consisting of continuous spinal anesthesia for hip fracture repair. Patients with continuous spinal anesthesia were more likely to receive regional analgesia (RA) via continuous femoral and lumbar paravertebral blocks, and these continuous blocks remained in situ an average of two postoperative days. Our group considers the use of RA in elderly patients to be opioid sparing and thus desirable, though we did not study this issue (11,12,13). The fact that there was no significant increase in morbidity or mortality associated with implementing RA for hip fracture repair (14) makes RA an attractive modality for perioperative pain control, especially important during the preoperative period. Neuraxial anesthesia, especially continuous spinal anesthesia should be considered an option for the geriatric population undergoing hip fracture surgery because of its ease of use and relative safety in addition to the fact that it offers optimal flexibility in the timing and duration of the surgery. The potential to decrease intraoperative hypotension (15) was also subjectively perceived to have been avoided.
Fragility hip fractures in the geriatric population can pose an economic burden to the health care system because this patient population is at risk for significant postoperative morbidity and mortality. Most published studies investigate perioperative parameters and their effects on postoperative morbidity and mortality with inconsistent results (16). Little data exists on the economic impact of the type of anesthesia used for hip surgery in the geriatric population. One study noted a lower cost of intraoperative anesthesia associated with using a single injection spinal anesthesia for intraoperative management. This study looked at only the intraoperative period, not at cost for the entire hospital stay (17). Our average estimated cost of hospital stay for hip fracture in the geriatric population is consistent with what has been published in the literature in the United States and similar to numbers published in Europe (2,17,18). Kates, et al. published a lower median cost of hospitalization than the current study ($11,417), but they did not add surgeon and geriatrician costs of care (19). The current study includes all estimated costs of hospitalization, including labor, procedures, and equipment.
Continuous spinal anesthesia was felt to provide good anesthesia with judicious dosing of local anesthetic to allow for improved hemodynamic control and avoidance of an invasive airway (14). It also provided optimal flexibility in scheduling, as patients with hip fractures are almost always unscheduled and emergency surgery. The timing of the end of surgery is usually not predictable, especially in a large very busy training hospital and continuous spinal anesthesia provided optimal timing flexibility and added little, or in the case of this study, no additional risk. Patients undergoing general anesthesia who failed tracheal extubation did so because of hemodynamic instability and poor postoperative ventilation, especially in the presence of preexistent pulmonary comorbidities. This study was thus performed on the premise that the patients who received RA would have lower incidences of postoperative ventilator requirements since they would require less systemic anesthetic and analgesics, and thus have reduced postoperative respiratory depression. Because ventilator dependence would necessitate postoperative ICU admission, avoidance of postoperative respiratory depression was believed to result in decreased cost of hospitalization, and an instinctive notion that these patients would have decreased morbidity and mortality. However, RA did not influence morbidity, in-patient mortality, or the number of patients who required postoperative ventilation via an endotracheal tube, which was low in both the GA and RA group. One patient in the RA group (1%) vs 10 patients in the GA group (4%) required postoperative mechanical ventilation. This makes it difficult to draw meaningful conclusions about this hypothesis. However, consistent with results from previous publications, this study finds that cost of hospitalization increased with delay in surgery. Intensive care unit admission expectedly also increased the cost of hospitalization.
This study differs from current literature in several ways. A recent Cochrane review of randomized controlled trials prescribing GA versus RA showed borderline improvement of short term mortality and decreased postoperative delirium with regional anesthesia, but no long term differences (20). Radcliff retrospectively reviewed a large database of Veterans Administration patients with hip fractures and showed increased short term mortality associated with intraoperative GA (versus RA). However, they did not investigate any role for peripheral nerve blockade for postoperative pain relief (21). Pedersen assessed the implementation of a pathway, which included early anesthesia evaluation, optimization of fluid and electrolyte balance, nutritional assessment and a social service consult in addition to postoperative pain management with a peripheral nerve block (22), and demonstrated a reduction in opioid requirements, postoperative complications and mortality. None of the cited studies used continuous intrathecal catheters for intraoperative anesthesia nor did they look at the financial implications of different treatments.
A weakness of this study includes its retrospective nature. Furthermore, the authors did not adjust the estimated cost of hospitalization data for inflation, but because the patient cohort was selected over a period of less than two years, it was felt that inflation would not alter the data significantly. Hospital charge data was collected instead of cost data, however, because cost-to-charge ratio for hospitals can be inferred using the cost-to-charge ratio from the Healthcare Cost and Utilization Project, this number was felt to adequately reflect cost of hospitalization of these patients. Data collected with regards to complications and comorbid conditions was based on ICD-9 codes, and therefore, the accuracy of the information collected is subject to the accuracy of the actual coding itself.
The fact that type of surgical anesthesia did not alter the patient’s cost of hospitalization or postoperative morbidity and mortality may be a result of the complexity of the patients’ medical history and the number of surgical and medical factors that all can contribute to the outcomes or it may simply reflect the natural coarse of life in this population, which consisted of 188 of 306 patients being older than 80 years old.
Conclusion
There is no difference in cost of hospitalization, re-hospitalization rates, postoperative morbidity, and in-patient mortality in geriatric patients, the majority being older than 80 years old, undergoing regional versus general anesthesia for hip fracture repair. Delay in surgery beyond 3 days and intensive care unit admission both increase cost of hospitalization in geriatric patients that receive surgery for hip fracture.
Acknowledgement
A special thanks to Christine Bono, PhD, Decision Support Services Analyst for the acquisition of financial data.
Footnotes
All authors have read and approved the manuscript. Dr. Boezaart is a consultant for Teleflex Medical and receives royalties for Arrow International for his invention of the “stimulating catheter”.
All other authors have no conflicts of interest to disclose.
References
- 1.International Osteoporosis Foundation webpage. Melton LJ., 3rd [accessed March 12, 2012];Hip fractures: a worldwide problem today and tomorrow. Bone. 1993 14(Suppl 1):S1–8. doi: 10.1016/8756-3282(93)90341-7. http://www.iofbonehealth.org/facts-and-statistics.html. [DOI] [PubMed] [Google Scholar]
- 2.Roudsari BS, Ebel BE, Corso PS, Molinari NA, Koepsell TD. The acute medical care costs of fall related injuries among the U.S. older adults. Injury. 2005;36:1316–22. doi: 10.1016/j.injury.2005.05.024. [DOI] [PubMed] [Google Scholar]
- 3.Forsen L, Meyer HE, Midthjell K, Edna TH. Survival after hip fracture: short- and longterm excess mortality according to age and gender. Osteoporos Int. 1999;10:73–8. doi: 10.1007/s001980050197. [DOI] [PubMed] [Google Scholar]
- 4.Center JR, Nguyen TV, Schneider D, Sambrook PN, Eisman JA. Mortality after all major types of osteoporotic fracture in men and women: an observational study. Lancet. 1999;353:878–82. doi: 10.1016/S0140-6736(98)09075-8. [DOI] [PubMed] [Google Scholar]
- 5.Haentjens P, Magaziner J, Colon-Emeric CS, Vanderschueren D, Millisen K, Velkeniers B, Bonnen S. Meta-analysis: excess mortality after hip fracture among older women and men. Ann Intern Med. 2010;152:380–90. doi: 10.1059/0003-4819-152-6-201003160-00008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Haleem S, Lutchman L, Mayahi R, Grice JE, Parker MJ. Mortality following hip fracture: trends and geographical variations over the last 40 years. Injury. 2008;39:1157–63. doi: 10.1016/j.injury.2008.03.022. [DOI] [PubMed] [Google Scholar]
- 7.Keene GS, Parker MJ, Pryor GA. Mortality and morbidity after hip fractures. Br Med J. 1993;307:1248–50. doi: 10.1136/bmj.307.6914.1248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Schroder HM, Erlandsen M. Age and sex as determinants of mortality after hip fracture: 3,895 patients followed for 2.5-18.5 years. J Orthop Trauma. 1993;7:525–31. doi: 10.1097/00005131-199312000-00007. [DOI] [PubMed] [Google Scholar]
- 9.Healthcare Cost and Utilization Project [accessed 09/09/2011];National Statistics. 2008 http://www.hcupus.ahrq.gov.
- 10.Leslie RS, Ghomrawi H. [assessed December 19, 2011];The Use of Propensity Scores and Instrumental Variable Methods to Adjust For Treatment Selection Bias. SAS Global Forum. 2008 Paper 366-2008. http://www2.sas.com/proceedings/forum2008/366-2008.pdf.
- 11.Gilbert TM, Hawkes WG, Hebel R, Hudson JI, Kenzora JE, Zimmerman SI, Felsenthal G, Magaziner J. Spinal anesthesia versus general anesthesia for hip fracture repair: a longitudinal observation of 741 elderly patients during 2-year follow-up. Am J Orthop. 2000;29:25–35. [PubMed] [Google Scholar]
- 12.Catro-Alves LJ, De Azevedo VL, De Freitas Braga TF, Goncalves AC, De Oliveira GS., Jr. The effect of neuraxial versus general anesthesia techniques on postoperative quality of recovery and analgesia after abdominal hysterectomy: a prospective, randomized, controlled trial. Anesth Analg. 2011;113:1480–6. doi: 10.1213/ANE.0b013e3182334d8b. [DOI] [PubMed] [Google Scholar]
- 13.Abbou-Setta AM, Beaupre LA, Rashiq S, Dryden DM, Hamm MP, Sadowski CA, Menon MR, Majumdar SR, Wilson DM, Karkhaneh M, Mousavi SS, Wong K, Tjosvold L, Jones CA. Comparative effectiveness of pain management interventions for hip fracture: a systematic review. Ann Intern Med. 2011;155:234–45. doi: 10.7326/0003-4819-155-4-201108160-00346. [DOI] [PubMed] [Google Scholar]
- 14.Foss NB, Kristensen BB, Bundgaard M, Bak M, Heiring C, Virkelyst C, Hougaard S, Kehlet H. Fascia iliaca compartment blockade for acute pain control in hip fracture patients: a randomized, placebo-controlled trial. Anesthesiology. 2007;106:773–8. doi: 10.1097/01.anes.0000264764.56544.d2. [DOI] [PubMed] [Google Scholar]
- 15.Minville V, Asehnoune K, Delussy A, Fourcade O, Colombani A, Rabinowitz A, Samii K. Hypotension during surgery for femoral neck fracture in elderly patients: effect of anaesthetic techniques. A retrospective study. Minerva Anestesiol. 2008;74:691–6. [PubMed] [Google Scholar]
- 16.Luger TJ, Kammerlander C, Gosch M, Luger MF, Kammerlander-Knauer U, Roth T, Kreutziger J. Neuraxial versus general anaesthesia in geriatric patients for hip fracture surgery: does it matter? Osteoporos Int. 2010;21(4 Suppl):S555–72. doi: 10.1007/s00198-010-1399-7. [DOI] [PubMed] [Google Scholar]
- 17.Chakladar A, White SM. Cost estimates of spinal versus general anaesthesia for fractured neck for femur surgery. Anaesthesia. 2010;65:810–4. doi: 10.1111/j.1365-2044.2010.06382.x. [DOI] [PubMed] [Google Scholar]
- 18.Sahota O, Morgan N, Moran CG. The direct cost of acute hip fracture care in care home residents in the UK. Osteoporos Int. 2011 doi: 10.1007/s00198-011-1651-9. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
- 19.Kates SL, Mendelson DA, Friedman SM. The value of an organized fracture program for the elderly: Early results. J Orthop Trauma. 2011;25:233–7. doi: 10.1097/BOT.0b013e3181e5e901. [DOI] [PubMed] [Google Scholar]
- 20.Parker MJ, Handoll HH, Griffiths R. Anaesthesia for hip fracture surgery in adults. Cochrane Database Syst Rev. 2004;4:CD00052. doi: 10.1002/14651858.CD000521.pub2. [DOI] [PubMed] [Google Scholar]
- 21.Radcliff TA, Henderson WG, Stoner TJ, Khuri SF, Dohm M, Hutt E. Patient risk factors, operative care, and outcomes among older community-dwelling male veterans with hip fracture. J Bone Joint Surg Am. 2008;90:34–42. doi: 10.2106/JBJS.G.00065. [DOI] [PubMed] [Google Scholar]
- 22.Pedersen SJ, Borgbjerg FM, Schousboe B, Pedersen BD, Jorgensen HL, Duus BR, Lauritzen JB. Hip Fracture Group of Bisperbjerg Hospital. A comprehensive hip fracture program reduces complication rates and mortality. J Am Geriatr Soc. 2008;56:1831–8. doi: 10.1111/j.1532-5415.2008.01945.x. [DOI] [PubMed] [Google Scholar]