The cost and consequences of proximal femoral fractures which require further surgery following initial fixation

Transcription

1 The cost and consequences of proximal femoral fractures which require further surgery following initial fixation C. Thakar, J. Alsousou, T. W. Hamilton, K. Willett From John Radcliffe Hospital, Oxford, United Kingdom We evaluated the cost and consequences of proximal femoral fractures requiring further surgery because of complications. The data were collected prospectively in a standard manner from all patients with a proximal femoral fracture presenting to the trauma unit at the John Radcliffe Hospital over a five-year period. The total cost of treatment for each patient was calculated by separating it into its various components. The risk factors for the complications that arose, the location of their discharge and the mortality rates for these patients were compared to those of a matched control group. There were 2360 proximal femoral fractures in 2257 patients, of which 144 (6.1%) required further surgery. The mean cost of treatment in patients with complications was ( to ), compared with 8610 ( to ) for uncomplicated cases (p ), with a mean length of stay of 62.8 (44.5 to 79.3) and 32.7 (23.8 to 35.0) days, respectively. The probability of mortality after one month in these cases was significantly higher than in the control group, with a mean survival of 209 days, compared with 496 days for the controls. Patients with complications were statistically less likely to return to their own home (p ). Greater awareness and understanding are required to minimise the complications of proximal femoral fractures and consequently their cost. C. Thakar, BSc, MBBS, MRCS, Orthopaedic Registrar J. Alsousou, LMSSA, MD, MRCS, Clinical DPhil Research Fellow T. W. Hamilton, BSc, MBChB, Academic Foundation Doctor K. Willett, FRCS, Professor of Orthopaedic Trauma Surgery, National Clinical Director for Trauma Care Kadoorie Centre for Critical Care Research and Education John Radcliffe Hospital, Headley Way, Headington, Oxford OX3 9DU, UK. Correspondence should be sent to Mr C. Thakar; chrish49@hotmail.com 2010 British Editorial Society of Bone and Joint Surgery doi: / x.92b $2.00 J Bone Joint Surg [Br] 2010;92-B: Received 12 April 2010; Accepted after revision 11 August 2010 Proximal femoral fractures continue to be the most common reason for admission to hospital following orthopaedic injury. 1 With an ageing population, the incidence of this fracture is predicted to rise. 2,3 An estimated hip fractures occur every year in the United Kingdom, with an annual cost of 1.7 billion to the National Health Service and social care services, which represents almost 2% of the annual budget of the NHS. 4 Fragility fractures of the hip in the elderly are a substantial cause of mortality and morbidity, 5-8 with approximately 10% of patients dying within 30 days and up to 27% within one year as a result of the fracture and associated comorbidities. 3 After surgery a significant increase in support for the activities of daily living (ADL) may be required, with many individuals having a greater dependence on social and community care. 9 This demand for limited resources is compounded where complications occur with treatment, or when more than one surgical procedure has been required. 9,10 Many studies have examined the morbidity, mortality and associated complications of proximal femoral fractures, 5-8 but few have assessed the financial cost of all complications which require surgery following the initial management The cost of treating an uncomplicated proximal femoral fracture in England has been reported to be between 7210 (2006) 10 and (2005), 15 with a mean duration of hospital stay of between 18 and 42 days. 11,16 The cost of treatment in complicated cases has been assessed as high as , 10 with a mean duration of stay of 109 days. 12 The majority of the costs arise from the extended duration of stay in hospital and the additional time in theatre. They may be calculated using the cost apportionment approach, as described by Hollingworth et al, 16 where the cost of treatment is broken down into its various components, including hotel costs, theatre costs, medical costs, ward costs, overheads and other expenses. The location to which the patient is discharged following treatment has also been investigated and a number of risk factors, such as age and male gender, which influence this, have been identified. 9,13 The impact of complications following treatment of these fractures is important in the current economic climate. Greater awareness and understanding are warranted. Recognition of risk factors may allow earlier intervention to reduce the rate or severity of complications and, in turn, the cost. VOL. 92-B, No. 12, DECEMBER

2 1670 C. THAKAR, J. ALSOUSOU, T. W. HAMILTON, K. WILLETT Table I. Social dependency and mobility scores Social dependency scale Score Description 1 Independent in domestic and social activities 2 Minimal help with shopping; may have meals on wheels 3 Dependent on social support up to three times a week 4 Dependent on social support more than three times a week, but less than daily. All domestic activities performed by spouse or carer 5 Dependent on social support more than once daily, or resident in a residential home. Personal care provided by spouse 6 Resident in nursing home or long-stay hospital Mobility score Score Description 0 Bedbound 1 Bed to chair 2 Mobilises with a frame and the assistance of two people 3 Mobilises with a frame and the assistance of one person 4 Mobilises with a frame alone, without the need for assistance 5 Walking stick 6 No aid The aim of this study was to calculate the additional hospital cost of further surgical interventions after the initial surgical management of patients with proximal femoral fractures. We assessed the potential risk factors for complications that arose in these patients compared with a matched control group. Their destination on discharge and the mortality rates were also compared. Patients and Methods This was a matched cohort study of all proximal femoral fractures presenting to our orthopaedic trauma unit between 1 March 2003 and 29 February A dedicated audit staff collected data prospectively for all patients with hip fractures. The collection was compatible with the National Hip Fracture Database and managed in accordance with the Caldicott principles. 17 There were 2360 proximal femoral fractures in 2257 patients. Of these fractures, 144 (6.1%) required one or more additional surgical procedure as a result of a complication of the primary operation. Further surgery was performed either during the initial admission in 53 patients or at a subsequent re-admission in 91. These cases were identified from the unit database and then cross-referenced with the operating theatre records and databases from radiology and microbiology. Deep surgical site infection was defined as microbiological confirmation of infection from culture of tissue samples taken deep to the fascia lata at a further operation. Using methods described previously, 10 each case was matched with two controls taken from the same prospectively collected dataset. The controls were matched based on eight factors (mean 7.6, range 6 to 8) known to influence outcome after proximal femoral fracture These were the same gender, age within four years, same fracture type (intra- or extracapsular), American Society of Anaesthesiologists (ASA) grade 21 within one grade, identical residence before fracture, an identical operation using parallel cancellous screws, a hemiarthroplasty or a total hip replacement (THR) for intracapsular fractures with a compression hip screw or intramedullary sliding hip screw for extracapsular fractures, social dependency to within one grade and mobility scores within one grade (Table I). The indications for fixation with parallel screws included a Garden type I or II fracture, or a young patient with a displaced intracapsular fracture. The patients were studied from their acute admission, intermediate care and/or community rehabilitation and until discharge. The total time spent in NHS care is termed superspell, which we divided into acute hospital stay and community hospital stay. The date and time of acute admission was recorded, as were the date, time, duration and grade of the operating surgeon for the primary procedure. Information including age, gender, comorbidities, ASA score, smoking status, mobility score, social dependency score and where patients were admitted from was also recorded. The details of any readmission to the trauma unit were noted. For those who sustained complications from their primary procedure, the type of complication was recorded and cross-referenced as described above. Information regarding mortality was obtained from the Office of National Statistics in November 2009, over one year after the last patient was included in the study. This allowed us to calculate the probability of mortality. Censored data have been accommodated in our statistical analysis using Kaplan-Meier survival analysis. The total cost of treatment for each patient was calculated by separating it into its constituent components and obtaining unit costs from the local finance department, as described by Hollingworth et al. 16 The figures quoted are based on costs between 2008 and 2009 in pounds sterling. THE JOURNAL OF BONE AND JOINT SURGERY

3 THE COST AND CONSEQUENCES OF PROXIMAL FEMORAL FRACTURES WHICH REQUIRE FURTHER SURGERY FOLLOWING INITIAL FIXATION 1671 Table II. Patient demographics of complicated cases and matched controls Complicated cases Matched controls p-value Mean age at fracture (yrs) (range) 82.6 (65 to 99) 83.0 (63 to 95) 0.57 Female (%) Mean ASA * grade Mean social dependency score (range) 2.8 (1 to 6) 2.8 (1 to 6) 0.98 Mean mobility score Type of fracture (%) Intracapsular 96 (66.6) Extracapsular 40 (27.7) Subtrochanteric 8 (5.6) Operation type (%) Parallel screws Compression hip screw Thompson Exeter trauma stem Total hip replacement Other Mean time to original fracture surgery (days) (range) 2.9 (0 to 9) 3.1 (0 to 7) 0.29 Cases performed by consultant (%) Cases performed by registrar (%) Mean operating time in minutes (range) 98.5 (30 to 195) 95.4 (20 to 275) 0.34 * ASA, American Society of Anaesthesiologists The hotel cost was added to supplementary departmental costs, which include theatre time, prosthetic costs, radiology and pharmaceuticals. A day spent on our unit was valued as 250 (2009) and the estimated cost of one hour of operating was 282. For those cases with confirmed deep infection the mean cost of a six-week course of intravenous antibiotics was added. This is our minimum requirement for all cases of implant-related infection. The primary outcome measure was the total cost of the acute care. Other outcome measures included the mortality and the length of stay in the acute or community hospital. The ultimate discharge destination (to the patient s home or to a residential or nursing home) was identified. Data on mortality were analysed at one and 12 months. Statistical analysis. All the data were analysed using SPSS statistics software version 17.0 (SPSS Inc., Chicago, Illinois), using a chi-squared test for independence with discrete data, a Mann-Whitney test with non-parametric and Student s t-test with parametric data. Direct logistical regression was used to identify risk factors for complications. The probabilities of survival and mortality were estimated using the Kaplan-Meier test and Cox s regression analysis. A p-value < 0.05 was considered significant. Results The 144 patients with complications requiring further surgery consisted of 118 women and 26 men, with a mean age of 82.6 years (65 to 99); 96 (66.7%) had sustained intracapsular fractures, 40 (27.8%) extracapsular fractures, and eight (5.6%) had subtrochanteric fractures. The intracapsular fractures were managed with a cemented Thompson hemiarthroplasty in 55 patients (57%), an Exeter Trauma Stem in 24 (25%), parallel cancellous screws in 15 patients (16%), a THR in one and a compression hip screw in one. The 40 extracapsular fractures were classified according to the number of constituent parts and were fixed according to the surgeon s preference. There were nine (22%) four-part fractures, of which seven were managed with a compression hip screw (one with a trochanteric plate), and two with a dynamic condylar screw (DCS). There were 19 (47%) with three-part fractures which were managed with a compression hip screw in 17 patients, two with trochanteric plates, and two with a DCS. The 12 (30%) with two-part fractures were treated with a compression hip screw. The eight subtrochanteric fractures were managed with a proximal femoral nail in three patients, a trochanteric femoral nail in one, a compression hip screw in three (two with trochanteric plates) and a DCS in one. There were 98 patients (68.1%) admitted from their own home, 20 (13.9%) from a residential care home or wardencontrolled flat, ten (6.9%) from nursing homes, 11 (7.6%) were already hospital in-patients at the time of the fracture, and five (3.5%) were admitted from other residential medical facilities. Table II outlines the demographics of the case and control groups with regard to their matched variables. There were no significant differences for these variables or pre-fracture residence between the two groups. Other potential confounding factors, including the time from admission to surgery, the percentage of cases with a VOL. 92-B, No. 12, DECEMBER 2010

4 1672 C. THAKAR, J. ALSOUSOU, T. W. HAMILTON, K. WILLETT Table III. Types of complication (%) Fixation failure (%) Superficial infection (%) Deep infection (%) Dislocation (%) Haematoma (%) Other (%) Screws 10 (27.2) (13.3) 3 (21.4) CHS * 21 (56.8) 6 (35.3) 7 (24.1) 0 1 (6.7) 4 (28.6) ETS 0 4 (23.5) 6 (20.7) 8 (25.0) 1 (6.7) 5 (35.7) Thompson 0 7 (41.2) 13 (44.8) 23 (71.9) 10 (66.7) 2 (14.3) THR (3.1) 0 0 Other 6 (16.2) 0 3 (10.3) 0 1 (6.7) 0 Total 37 (25.7) 17 (11.8) 29 (20.1) 32 (22.2) 15 (10.4) 14 (9.7) * CHS, compression hip screw ETS, Exeter trauma stem THR, total hip replacement Table IV. Results of bacteriological culture Organism Mean number (%) Patients complicated by deep infection Staphylococcus aureus 9 (31.0) Coagulase-negative staphylococcus 7 (24.1) MRSA * 10 (34.5) E. coli 1 (3.4) Other 2 (6.9) Patients complicated by superficial infection Staphylococcus aureus 6 (35.3) Coagulase-negative staphylococcus 9 (47.1) MRSA 1 (5.9) E. coli 2 (11.8) * MRSA, methicillin-resistant Staphylococcus aureus consultant as the lead surgeon and the mean operating time, are outlined in Table II. No significant difference was found between these variables. The types of complication by procedure are outlined in Table III. Other complications included avascular necrosis (AVN), peri-prosthetic fracture, cement in the acetabulum, ongoing pain, wound dehiscence, wound discharge and displacement of the greater trochanter. The operative management of these complications varied according to factors affecting both the patient and the surgical procedure. All cases of suspected infection or a discharging haematoma had at least one exploration of the wound with debridement and microbiological sampling. Further management depended on the intra-operative findings and the microbiology cultures. The total number of washouts required for cases complicated by infection or haematoma was 92 in 61 patients, with one patient requiring four procedures. The mean number of washouts for patients complicated by infection was 1.45 (one to four) and for those with a haematoma 1.06 (one to two). Infection was the complication in 46 cases (31.9%). Our management of suspected deep wound infection consists of full and, if necessary, repeated surgical debridement, together with close liaison with the clinical microbiologists and the bone infection unit. Deep infection was confirmed in 29 cases (20.1%) by positive culture of tissue samples taken deep to the fascia lata. These patients received intravenous antibiotics for six weeks, followed by oral suppressive therapy as appropriate. The infecting organisms for deep and superficial infections are shown in Table IV. The mean number of debridements performed was 1.6 (one to four) for deep infection and 1.2 (one to two) for superficial infection. Of the 29 cases of confirmed deep infection, 13 (44.8%) required removal of their implants: 11 had a Girdlestone arthroplasty with removal of the implant and cement with or without an antibiotic cement spacer, one had a primary exchange of implants, and another a twostage revision to a THR. Of 32 dislocated prostheses, three had a primary Girdlestone procedure, two were converted to a THR, one patient died before revision, and the remaining 26 patients had closed reduction of the prosthesis. Of these 26, 18 required further surgery for a subsequent dislocation and were managed with either a further reduction, a Girdlestone procedure or a THR; one patient required a total of six operations. The detailed analysis of the length of stay by complication, further surgery requirement and the controls is shown in Table V. For the purposes of analysis, where there was more than one admission the lengths of stay have been summated. The destination following discharge was considered to be that after the final admission. The 30-day difference in the mean total time spent in NHS care was significant (p ), the majority being due to the increase in acute hospital bed days (mean increase 29 days, p ) rather than in the community hospital. Figure 1 shows the trend for this period for both cases and controls. The mean length for the complicated cases is significantly higher than that of the controls, being 62.8 days as opposed to 32.7 (p = 0.02). When analysing mortality, no significant difference was seen between cases and controls in the first three months. However, after that time the probability of mortality in complicated cases was significantly higher than in the matched control group (Fig. 2). The Kaplan-Meier test showed that the mean survival of complicated cases was 209 days, as opposed to 496 days for the matched controls (p = 0.035), with median values of 1118 days and 1659 days, respectively (95% confidence interval (CI) for THE JOURNAL OF BONE AND JOINT SURGERY

5 THE COST AND CONSEQUENCES OF PROXIMAL FEMORAL FRACTURES WHICH REQUIRE FURTHER SURGERY FOLLOWING INITIAL FIXATION 1673 Table V. Mean duration of in-patient stay (days) Acute stay (range) Community stay (range) Total NHS stay (range) Cases Case-matched controls p-value Cases Case-matched controls p-value Cases Case-matched controls p-value Fixation failure 39.5 (12 to 93) 14.7 (2 to 56) 40.9 (9 to 99) 38.5 (0 to 125) (12 to 143) 30.0 (4 to 132) 0.02 Superficial infection 47.0 (19 to 134) 14.8 (3 to 34) 29.2 (2 to 93) 42.6 (0 to 101) (19 to 170) 35.0 (3 to 116) 0.15 Deep infection 62.0 (32 to 137) 14.8 (3 to 38) 35.9 (7 to 107) 50.4 (0 to 167) (33 to 237) 34.3 (3 to 193) 0.26 Haematoma 33.8 (21 to 48) 13.2 (6 to 38) (1 to 70) 32.0 (0 to 64) (21 to 104) 23.8 (8 to 90) 0.82 Dislocation 43.4 (12 to 94) 17.3 (6 to 41) 43.3 (1 to 174) 40.4 (0 to 173) (15 to 209) 34.2 (6 to 198) 0.11 Overall 44.3 (5 to 137) 15.3 (1 to 56) 36.5 (1 to 174) 43.0 (0 to 173) (5 to 237) 32.7 (1 to 237) 0.02 Patients (%) Case superstay Control superstay Time (days) Fig. 1 Graph showing the trend line for the total time spent in NHS care for both cases and controls. complicated cases to , and for matched controls to ). Cox s regression analysis of covariance confirmed that all complications increased the probability of mortality (p = 0.001) apart from superficial infection (p = 0.12) (Fig. 3). At the final discharge those patients who had sustained complications were statistically less likely to return to their own home (p = 0.01) and more likely to be referred to continuing care (p ). Overall, only 46.5% of complicated cases (67 patients) ultimately returned to their beforefracture residence, compared to 54.9% of matched controls. This difference was not significant (p = 0.12) (Table VI). A Kolmogorov-Smirnov test was performed for normality and gave values < 0.05, and so a Mann-Whitney non-parametric test was used to compare costs. The mean financial cost of treatment of the complicated cases was per patient ( to ), more than twice as much as for the matched controls 8610 ( to , p ). This figure includes those who had died early with lower mean costs (Table VII and Fig. 4). Of the difference in costs, 72% was due to the extended stay in the acute trauma unit and 20% due to the cost of additional procedures. Direct logistical regression analysis was performed to identify risk factors predictive of the development of complications following the initial primary surgery. The model contained 15 independent variables, listed in Table VIII. The full model containing all factors was statistically significant (chi-squared (15 degrees freedom, n = 432 (144 cases, 288 controls)) = 36, p = 0.02), indicating that it was able to distinguish between controls and cases with complications. The model as a whole explained between 8.1% (Cox and Snell R 2 ) and 11.2% (Nagelkerke R 2 ) of the cases, and correctly identified 69.0% of complicated cases. As shown in Table VIII only two of the independent variables made a unique, statistically significant contribution to the model. These were the need for post-operative blood transfusion and the use of enteral steroids. Receiving a post-operative blood transfusion had an odds ratio (OR) for complications of 2.54, whereas enteral steroids had an OR of Subgroup analysis was performed to identify risk factors specific to the relatively common complications of failure of fixation, superficial and deep infection, and dislocation. The use of steroids was predictive of failure of fixation, OR = 8.99 (1.06 to 76.1, p = 0.04). For deep infection the risk factors were dementia (OR = 6.72, 1.16 to 38.8, p = 0.03), smoking (OR = 23.3, 1.45 to , p = 0.03) and post-operative blood transfusion (OR = 33.13, 5.69 to , p ). In superficial infection a history of a VOL. 92-B, No. 12, DECEMBER 2010

6 1674 C. THAKAR, J. ALSOUSOU, T. W. HAMILTON, K. WILLETT Probability of mortality Case Control Probability of mortality Complication Wound dehiscence Superficial infection Deep infection Dislocation Failure of fixation Haematoma Pain Time from fracture to death (days) Time from fracture to death (days) Fig. 2 Fig. 3a Kaplan-Meier graph showing the probability of mortality in cases and controls over the study period. cerebrovascular accident (OR = 22.93, to , p = 0.03), diabetes (OR = 28.42, 1.27 to , p = 0.04) or post-operative blood transfusion (OR = 28.48, 2.47 to , p ) were predictive. For dislocation, the operation not being performed by a consultant (OR = 5.46, 1.14 to 26.32, p = 0.03) and post-operative blood transfusion (OR = 7.33, 2.16 to 24.9, p = 0.01) were linked to the development of this complication. Caution must be used when interpreting these results in the light of the small number of patients taking steroids (seven), smoking (six), having a history of a cerebrovascular accident (seven) or diabetes (seven). Probability of mortality Complication Wound dehiscence Superficial infection Deep infection Dislocation Failure of fixation Haematoma Pain Discussion In older people hip fractures are associated with a substantial level of morbidity and mortality. 3,5-8 Complications following surgical fixation have been linked to a higher degree of mortality, duration of hospital stay, cost and social dependence. Few studies have examined the financial cost and consequences of the complications that may follow such surgery. Edwards et al 12 and Pollard et al 10 studied the impact of infection in terms of cost and outcome following fixation of these fractures. Both demonstrated statistically significant differences in the cost of treatment and the length of stay. Our study supports these findings in reoperation for all causes, as well as demonstrating increasing risk of mortality at one year. Our rate of deep infection for the study period was 1.2%, which is comparable to other reported rates of Time from fracture to death (days) Fig. 3b Graphs showing Cox s regression analysis of the effect of complications on probability of mortality for (a) the controls, (b) the cases. between 1.3% 22 and 3.6%. 23 Methicillin-resistant Staphylococcus aureus (MRSA) was found to be the most common organism. The mean length of acute stay for those patients who had a deep infection was 62 days (3 to 137), compared with a mean of 15 days (3 to 38) in the control group. Deep infection was the most costly complication, with a mean cost of ( to ). This finding is consistent with a previous study 10 in our unit THE JOURNAL OF BONE AND JOINT SURGERY

7 THE COST AND CONSEQUENCES OF PROXIMAL FEMORAL FRACTURES WHICH REQUIRE FURTHER SURGERY FOLLOWING INITIAL FIXATION 1675 Table VI. Patient destination after discharge Complicated cases (%) Matched controls (%) p-value Own home Residential hospital Nursing home Died Acute hospital Rehabilitation Other Returning to original residence Table VII. Cost of overall care ( (range)) per patient, by complication) Fixation failure Superficial infection Deep infection Acute stay Community stay Total Cases Controls p-value Cases Controls p-value Cases Controls p-value ( to ) ( to ) ( to ) ( to ) (0 to 21780) (0 to 20460) ( to ) ( to ) (0 to 23540) Haematoma ( to ) ( to ) Dislocation ( to ) ( to ) Overall ( to ) ( to ) (0 to 15400) (0 to 38280) (0 to 38280) (0 to 4400) (0 to 3674) ( to ) ( to ) ( to ) ( to ) ( to ) ( to ) ( to ) ( to ) ( to ) ( to ) ( to ) ( to ) (0 to 23100) (0 to 15180) (0 to 38060) (0 to 38060) which found the mean cost of an infection to be , and another study by Edwards et al 12 that calculated the mean cost to be All complications associated with re-operation in our patients were associated with an increase in cost and length of stay compared to the matched control group, with a mean cost of ( to ) (p ) and mean length of acute stay of 44.3 days (5 to 137) (p ). Deep infection had the greatest impact. Infection was the most common complication requiring subsequent surgical intervention (32%), followed by failure of fixation (25.7%). Those patients who had sustained complications of the initial procedure requiring further surgery were statistically less likely to be discharged to their own home (p = 0.01) and more likely to require continuing rehabilitation (p ) following a community hospital stay. After extended hospitalisation, only 46.5% of patients with complications returned to their original residence, compared to 54.9% of matched controls; although this difference was not significant (p = 0.12). The strongest predictor of post-operative complications requiring further surgery was the need for post-operative transfusion, with an OR of Analysis of our results also suggests a possible increased risk in those patients taking steroids and those with dementia. However, these results should be interpreted with caution owing to the small numbers. The strengths of this study are its size and the matching of eight variables that represent a valid comparison between the two groups. Using 2:1 matching increased the power of the study. The continued follow-up of all patients until discharge from hospital care ensured the accuracy of the study. Although we acknowledge limitations in determining the accuracy of the costs, the majority arise from hospital stay, theatre time and the price of the implant, all of which we were able to determine from our local departments. Taking an average cost of a six-week course of antibiotics for those patients complicated by deep infection ensured a standard cost for each patient. As proximal femoral fractures are the leading cause of admission for trauma, prevention of complications is vital in order to reduce the cost to both patients and society. Our study has shown the financial cost following complications to be over twice as much as that of controls. Although our calculations have accounted for the majority of direct costs, they do not account for the indirect effect on society with regard to employment, and the financial impact on relatives and carers. Although our study did not reveal any variable that might be corrected to prevent a complication, it did highlight an increased risk with the need for post-operative transfusion. We advise closer observation and increased awareness of patients who require post-operative transfusion, so as to detect any complication earlier and thereby reduce morbidity and mortality. This has been observed in other specialties, such as cardiothoracic surgery. 24 Although this should be standard practice, increasing awareness in these groups of patients would help to deliver a higher quality of care as well as poten- VOL. 92-B, No. 12, DECEMBER 2010

8 1676 C. THAKAR, J. ALSOUSOU, T. W. HAMILTON, K. WILLETT Pain Avascular necrosis Control Case Haematoma Superficial infection Dislocation of prosthesis Failure of fixation Deep infection Fig. 4 Graph showing the mean cost ( ) per patient of treatment in each complication category. Table VIII. Analysis of risk factors for the development of complications following the initial primary surgery for proximal femoral fractures Risk factor Odds ratio 95% CI * β coefficient p-value Cardiovascular disease to Stroke to Respiratory disease to Renal disease to Diabetes to Rheumatoid disease to Dementia to Parkinson s disease to Malignancy to Smoking to Enteral steroids to Warfarin to Consultant as surgeon to Transfused pre-operatively to Transfused post-operatively to * CI, confidence interval tially reducing costs. Although transfusion has been observed to cause a reduction in the recipient s immune response, 25 and hence an increased risk of infection, we did not note a significant difference in the rates of infection between those who were transfused and those who were not in the complicated cases. However, when the cohort of all the patients in our study was analysed there was a statistical increase in both superficial and deep infection in those who received a post-operative transfusion, compared to those who did not. We demonstrated a statistically significant increase in infection rates between those patients who underwent arthroplasty and those who did not (p ). There was no statistically significant difference between operating times in these two groups (p = 0.84) and no other significant variable to account for this. We therefore assume that the difference is due to the greater degree of soft-tissue trauma involved with arthroplasty. As is commonly appreciated, rarely is one single factor responsible for infection. However, our study does suggest that this is the most important complication to avoid, as it was the most common (31.9%) and the most costly. Surgical site infection is a major problem, but more so when metalwork is involved. All procedures were performed in the same theatre, under the same conditions using the same precautions and sterility. Minimising THE JOURNAL OF BONE AND JOINT SURGERY

9 THE COST AND CONSEQUENCES OF PROXIMAL FEMORAL FRACTURES WHICH REQUIRE FURTHER SURGERY FOLLOWING INITIAL FIXATION 1677 this complication by optimising patients medically, with appropriate administration of antibiotics and careful handling of soft tissue, together with post-operative wound management, is extremely important. 26 This study reinforces the need to optimise the pre-operative care of patients with proximal femoral fractures in order to reduce the complication rate and the associated costs of health and social care. Supplementary material A further opinion by Dr. T. Horgervorst is available with the electronic version of this article on our website at Listen live Listen to the abstract of this article at No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. References 1. Parker M, Johanson A. Clinical review: hip fracture. BMJ 2006;333: Johnell O, Gullberg B, Allander E, Kanis JA. The apparent incidence of hip fracture in Europe: a study of national register sources. MEDOS Study Group. Osteoporos Int 1992;2: Hollingworth W, Todd CJ, Parker MJ. The cost of treating hip fractures in the twenty-first century: short report. Osteoporos Int 1996;6(Suppl 2): No authors listed. Clinical practice guideline for the assessment and prevention of falls in older people. NICE Clinical Guideline 21, (date last accessed 11 August 2010). 5. Currie CT. Hip fractures in the elderly: beyond the metalwork. BMJ 1989;298: Magaziner J, Simonsick EM, Kashner TM, Hebel JR, Kenzora JE. Predictors of functional recovery one year following hospital discharge for hip fracture: a prospective study. J Gerontol 1990;45:M Zuckerman JD. Hip fracture. N Engl J Med 1996;334: Bentler SE, Liu L, Obrizan M, et al. The aftermath of hip fracture: discharge placement, functional status change, and mortality. Am J Epidemiol 2009;170: Greendale GA, Barrett-Connor E, Ingles S, Haile R. Late physical and functional effects of osteoporotic fracture in women: the Rancho Bernardo Study. J Am Geriatr Soc 1995;43: Pollard TC, Newman JE, Barlow NJ, Price JD, Willett KM. Deep wound infection after proximal femoral fracture: consequences and costs. J Hosp Infect 2006;63: Palmer SJ, Parker MJ, Hollingworth W. The cost and implications of reoperation after surgery for fracture of the hip. J Bone Joint Surg [Br] 2000;82-B: Edwards C, Counsell A, Boulton C, Moran CG. Early infection after hip fracture surgery: risk factors, costs and outcome. J Bone Joint Surg [Br] 2008;90- B: Deakin DE, Wenn RT, Moran CG. Factors influencing discharge location following hip fracture. Injury 2008;39: Haentjens P, Autier P, Barette M, Boonen S; Belgian Hip Fracture Study Group. The economic cost of hip fractures among elderly women: a one-year, prospective, observational cohort study with matched-pair analysis. J Bone Joint Surg [Am] 2001;83-A: Lawrence TM, White CT, Wenn R, Moran CG. The current hospital costs of treating hip fractures. Injury 2005;36: Hollingworth W, Todd C, Parker M, Roberts JA, Williams R. Cost analysis of early discharge after hip fracture. BMJ 1993;307: No authors listed. The Caldicott Committee: report on the review of patient-identifiable information. Department of Health, (date last accessed 11 August 2010). 18. Ions GK, Stevens J. Prediction of survival in patients with femoral neck fractures. J Bone Joint Surg [Br] 1987;69-B: Hubble M, Little C, Prothero D, Bannister G. Predicting the prognosis after proximal femoral fracture. Ann R Coll Surg Engl 1995;77: Ryder SA, Reynolds F, Bannister GC. Refining the indications for surgery after proximal femoral fracture. Injury 2001;32: Fehrenbach MJ. ASA Physical Status Classification System. (date last accessed 5 October 2010). 22. Partanen J, Syrjälä H, Vähänikkilä H, Jalovaara P. Impact of deep infection after hip fracture surgery on function and mortality. J Hosp Infect 2006;62: Mackay DC, Harrison WJ, Bates JH, Dickenson D. Audit of deep wound infection following hip fracture surgery. J R Coll Surg Edinb 2000;45: Scott BH, Seifert FC, Grimson R. Blood transfusion is associated with increased resource utilisation, morbidity and mortality in cardiac surgery. Ann Card Anaesth 2008;11: Blajchman MA. Immunomodulation and blood transfusion. Am J Ther 2002;9: Quinn A, Hill AD, Humphreys H. Evolving issues in the prevention of surgical site infections. Surgeon 2009;7: VOL. 92-B, No. 12, DECEMBER 2010