OUP user menu

Effects of socioeconomic position on 30-day mortality and wait for surgery after hip fracture

Anna Patrizia Barone, Danilo Fusco, Paola Colais, Mariangela D'Ovidio, Valeria Belleudi, Nera Agabiti, Chiara Sorge, Marina Davoli, Carlo Alberto Perucci
DOI: http://dx.doi.org/10.1093/intqhc/mzp046 379-386 First published online: 19 October 2009


Objective In countries where the National Health Service provides universal health coverage, socioeconomic position should not influence the quality of health care. We examined whether socioeconomic position plays a role in short-term mortality and waiting time for surgery after hip fracture.

Design Retrospective cohort study.

Settings and participants From the Hospital Information System database, we selected all patients, aged at least 65 years and admitted to acute care hospitals in Rome for a hip fracture between 1 January 2006 and 30 November 2007. The socioeconomic position of each individual was obtained using a city-specific index of socioeconomic variables based on the individual's census tract of residence.

Main outcome measures Three different outcomes were defined: waiting times for surgery, mortality within 30 days and intervention within 48 h of hospital arrival for hip fracture. We used a logistic regression to estimate 30-day mortality and a Cox proportional hazard model to calculate hazard ratios of intervention within 48 h. Median waiting times were estimated by adjusted Kaplan–Meyer curves. Analyses were adjusted for age, gender and coexisting medical conditions.

Results Low socioeconomic level was significantly associated with higher risk of mortality [adjusted relative risk (RR) = 1.51; P < 0.05] and lower risk of early intervention (adjusted RR = 0.32; P < 0.001). Socioeconomic level had also an effect on waiting times within 30 days.

Conclusions Individuals living in disadvantaged census tracts had poorer prognoses and were less likely than more affluent people to be treated according to clinical guidelines despite universal healthcare coverage.

  • hip fracture
  • inequalities
  • information systems
  • mortality
  • surgery


Hip fractures are the leading cause of hospitalization for injuries among the elderly population and have substantial impact on both the patient and the healthcare system. Many studies have explored the associations among patient characteristics, treatment processes, time to surgery and different outcomes in patients hospitalized for hip fracture [1, 2]. However, only a few studies have evaluated the impact of socioeconomic position on patient outcomes or on timing of hip-fracture surgery and the results have been discordant. Studies have shown no substantial effect of socioeconomic variables on inpatient mortality [3] or waiting times for surgery [35]; in contrast, other studies found a higher level of socioeconomic deprivation is associated with higher rates of operative delay [6] and higher probability of negative outcomes after hip fracture [6, 7]. Moreover, findings vary in different countries. For example, income does not appear to affect the probability of in-hospital mortality in the USA and Canada; but, in the former, patients living in higher-income areas tend to move up the queue for hip-fracture surgery more quickly than lower-income persons [8].

The relationship between socioeconomic position and health is complex and influenced by various factors, including the characteristics of healthcare systems. Healthcare disparities have been highlighted in the USA [9], UK [10] and many Western countries [11, 12]. In Italy, where the National Health Service (NHS) provides universal health coverage, economic or social barriers are not expected; yet, inequalities in obtaining optimal care have recently been reported [1315]. However, disparities in the quality of health care for acute hip fracture have not been documented to date. The present study was conducted to provide a new contribution examining the effect of socioeconomic position on short-term mortality and waiting time for surgery after hip fracture.


Design and data sources

We conducted a retrospective cohort study of patients residing in Rome (2 705 603 resident inhabitants as of January 2007 [16]). The Hospital Information System (HIS) database was used to identify cases of hip fracture, the selected outcomes and coexisting patient medical conditions. Hospital discharge abstracts included information on patients' characteristics, discharge diagnoses (up to 6) and procedure codes (up to 6) according to the International Classification of disease, Ninth Revision, Clinical Modification (ICD-9-CM).

We also used additional information from the Healthcare Emergency Information System (HEIS) database to increase the probability of identifying patients' comorbidities and to calculate time to death and time to surgery from the actual arrival at hospital [admission to Emergency Department (ED) or hospital ward]. The HEIS collects all ED-visit records in the Lazio region (a central Italy region whose capital city is Rome) including information on patients' characteristics, four categories of patient severity based on triage (red, yellow, green and white), main diseases, some clinical parameters, performed treatments and diagnoses at discharge (most information is codified by ICD-9-CM codes).

Deaths during the study period were identified using both the HIS (discharge disposition: death) and the Mortality Information System (MIS) of Lazio. The MIS database includes information on demographic characteristics (name, age, gender, fiscal code, place and date of birth, residence, marital status and occupation), as well as date, place and cause of death (codified by ICD-9 codes). HIS records were linked with HEIS and MIS records.

Study population

From the HIS database, we selected all patients with the following criteria: aged at least 65 years, resident of Rome and admitted to an acute care hospital in Rome for a hip fracture (ICD-9-CM diagnosis codes 820.0–820.9 in any position) between 1 January 2006 and 30 November 2007.

We excluded admissions of patients if they met any of the following criteria:

  1. hospitalized for hip fracture in the previous 2 years,

  2. transferred from other acute care hospitals or EDs (patients admitted to a given ED or hospital for hip fracture and coded as ‘transferred from’ other not identified acute care facility or ED),

  3. had multiple significant trauma (DRGs 484–487),

  4. directly admitted to intensive care units,

  5. died within 48 h of admission without intervention (patients who could have not been operated on due to poor baseline clinical conditions) and

  6. had a principal or secondary diagnoses of malignant neoplasms (codes 140.0–208.9) in the index admission (current admission for hip fracture) or in previous hospitalizations during the last 2 years.


We defined three different outcomes: mortality within 30 days of hospital arrival for hip fracture, waiting time for surgery and intervention within 48 h (0–1 day) of hospital arrival. Date of hospital arrival corresponded to the date of the index or ED admission. The interventions were identified by the following ICD-9-CM codes: total or partial hip replacement (codes 81.51, 81.52) and reduction of fracture (codes 79.00, 79.05, 79.10, 79.15, 79.20, 79.25, 79.30, 79.35, 79.40, 79.45, 79.50, 79.55).

Index of socioeconomic position

Socioeconomic position was the primary independent variable. We used a city-specific index, constructed for Rome, based on census data aggregated at the level of the census tract of residence. Data representing various dimensions of deprivation (education, occupation, crowding, immigration, family composition and home ownership) were provided by ISTAT, the Italian National Institute of Statistics. For each dimension of deprivation a variety of indicators were considered. A factorial analysis was performed in order to create a composite index of socioeconomic position. The indicators were weighted for the population of the census tract and were algebraically combined by using the factorial scores as weights. The city-specific index was assigned to patients on the basis of location of residence at discharge. To obtain categorical values for the aggregated indicator of socioeconomic position, the quintiles from the distribution by census tract were calculated. Detailed construction of this index has been described elsewhere [1719]. In our study, we combined quintiles 2–4 to highlight the differences between the extreme categories as the middle ones may be misclassified. Therefore, we defined three levels of socioeconomic position: I (high), II (intermediate) and III (low).

Coexisting medical conditions

Chronic comorbidities and/or severity characteristics, potentially associated with the outcomes under study, were chosen among the conditions identified in the literature [1, 2, 6, 8] and those empirically tested in the Mattoni-outcome project funded by the Italian Ministry of Health [20]. We defined potential risk factors as diseases of central nervous system, cardiac and vascular/cerebrovascular disorders, diabetes, hypertension, chronic obstructive pulmonary disease (COPD), chronic renal disease, liver disease, nutritional deficiencies and other conditions that could play a role in the outcomes under study. The potential risk factors were identified on the basis of ICD-9-CM codes registered either in the index hospitalization or in previous hospital or ED admissions during the last 2 years. Acute events that occurred during the index admission, which could be complications of care, were not included. Details and ICD-9-CM codes are reported in the Appendix.

Statistical analysis

Crude 30-day mortality rates, median waiting times for hip surgery and proportion of interventions within 48 h were calculated. Proportions of interventions within 48 h were defined as the number of patients (aged at least 65 years) who underwent intervention within 48 h out of the number of patients (aged at least 65 years) hospitalized for hip fracture. The median wait times until surgery were estimated by calculating the number of days elapsed between the date of hospital arrival and the date of hip-fracture surgery. Waiting times greater than 30 days were set at 30 since the most of interventions were performed within 30 days of hospital arrival.

We used multivariate regression analysis to assess the effect of socioeconomic position on 30-day mortality and chance of intervention within 48 h of hospital arrival, controlling for variations in other factors (age, gender and coexisting medical conditions) that could affect the outcomes under study. Among all factors potentially associated with the outcomes under study, age and gender were considered as a priori risk factors; the others were selected by a stepwise bootstrap procedure. In this approach, 100 replicated samples of 1000 observations were selected with replacement from the original data set. A stepwise procedure (significance for input 0.10 and removal 0.05) was applied to each replicated sample and only the risk factors selected in at least 50% of the procedures were included in the predictive models. Comorbidities were divided for conditions registered during the index hospitalization and previous hospital or ED admissions.

We used a logistic regression to estimate 30-day mortality and a Cox proportional hazard model to calculate hazard ratios (HRs) of intervention within 48 h. To estimate median waiting times for surgery by socioeconomic position, adjusted Kaplan–Meyer curves were calculated, as the proportionality condition was not verified. Odds ratios (ORs) were converted into relative risks (RRs) by the following formula: Embedded Image

Level I of the city-specific index (high socioeconomic position) was the reference category. Since conventional regression analysis does not account for lack of independence across levels of nested data (i.e. individuals nested within hospitals), we also used multilevel analysis as a sensitivity analysis, with hospitals as the level of aggregation. We assumed that the hospital-level data were normally distributed as the number of admissions by hospital was higher than 150. In addition, we verified the assumption of no correlation between any hospital-level disturbance term and socioeconomic position, using the likelihood ratio test. The level of significance was set at 5% (P < 0.05) and all analyses were undertaken using SAS Version 8.2 [21].


We studied 5051 admissions for hip fracture in Rome from 2006 to 2007. Overall, 88.2% of hip fractures were treated by total/partial hip replacement or reduction of fracture, 6.3% of patients were operated on within 48 h, and the median preoperative waiting time was 6 days. Of the 4454 patients who underwent surgery, 4137 (92.9%) waited for more than 48 h. The 30-day mortality was 6.1%.

The majority of patients were women (78.2%) and the mean age was 82.4 years (SD = 7.3) (Table 1). The proportion of males who did not undergo surgery was higher than that of females, 16.3% and 10.6%, respectively. Overall, people with comorbidities tended to have higher risks of mortality and lower risk of intervention within 48 h than individuals without coexisting medical conditions (Table 1). In particular, the risk of dying within 30 days was more than three times higher in patients with nutritional deficiencies and chronic renal disease, whereas the probability of early intervention was 5.5 times lower when the latter condition was present. Individuals with ischemic heart disease and COPD experienced the longest waiting times (median: 8 days) (Table 1).

View this table:
Table 1

Characteristics of people admitted for hip fracture

Risk factorAdmissions (n)Mortality (%)Relative riskP-valueIntervention within 48 h (%)HRP-valueMedian waiting time (days)
Age (mean: 82.4, SD: 7.3)
 Male (reference)110211.416.41
Presence of coexisting conditions
 Diseases of central nervous system (dementia, Parkinson, hemiplegia)5779.21.610.0024.50.690.076
 Ischemic heart disease6777.71.310.082.50.370.0008
 Other cardiac disorders (including heart failure)7509.91.810.0003.50.510.0016
 Circulatory disorders7859.21.650.0003.80.570.0036
 Chronic renal disease25019.63.620.0001.20.180.0036
 Other chronic disease (liver, pancreas, intestine)13113.72.320.0015.30.850.676
 Bone and joint disorders2577.01.150.565.80.930.775
 Nutritional deficiencies3420.63.420.0012.90.470.454
 Blood disorders11286.71.110.413.70.530.0006

Older age was associated with slightly higher risk of death (adjusted RR = 1.1; P < 0.001), but not with the probability of an early-surgical approach (adjusted HR = 1.0; P: 0.649). Risk of death was lower among women (adjusted RR = 0.44; P < 0.001), but both men and women had the same probability of undergoing surgery within 48 h (adjusted HR = 0.93; P = 0.61) (data not shown).

Table 2 reports the number of ordinary admissions for hip fracture, death rates, crude and adjusted RRs and their statistical significance, by the level of socioeconomic position. We found that lower socioeconomic position (level III) was significantly associated with higher risk of mortality (crude RR = 1.55; P < 0.05), even after controlling for patients characteristics such as age, gender and coexisting medical conditions. Individuals living in disadvantaged tracts were 50% more likely to die within 30 days than individuals in level I (adjusted RR = 1.51; P < 0.05) (Table 2). As shown in Table 3, people in socioeconomic level III were three times less likely to undergo intervention within 48 h than more privileged persons (adjusted RR = 0.32; P < 0.001). Small differences were found in the adjusted waiting times within 30 days, ranging from 5 days for the individuals in level I to 7 days in the level III (Table 4). We presented only the results from the classical approach, as they were similar to those obtained by applying the multilevel approach. We found no correlation between socioeconomic position and any hospital-level disturbance term for the three outcomes under study.

View this table:
Table 2

Thirty-day mortality rate and crude and adjusted relative risks by socioeconomic position

Socioeconomic positionAdmissions (n)30-day mortality rate (%)Crude relative riskP-valueAdjusted relative riskaP-value
I (high)11875.011
II (intermediate)31226.21.240.1431.240.14
III (low)7427.71.550.0191.510.03
  • aAdjusted for: age, gender, COPD, diabetes, cerebrovascular diseases, cerebrovascular diseases current admission (CA), dementias, dementias CA, chronic renal diseases, chronic renal diseases CA.

View this table:
Table 3

Proportion of interventions within 48 h and crude and adjusted HRs by socioeconomic position

Socioeconomic positionAdmissions (n)Interventions within 48 h (%)Crude HRP-valueAdjusted HRaP-value
I (high)11879.011
II (intermediate)31226.10.660.0010.690.002
III (low)7422.80.310.0000.320.000
  • aAdjusted for: age, gender, COPD, other forms of chronic ischemic heart diseases, blood disorders, blood disorders current admission (CA), cerebrovascular diseases, cerebrovascular diseases CA.

View this table:
Table 4

Crude and adjusted waiting times for surgery by socioeconomic position

Socioeconomic positionAdmissions (n)Interventions (%)Median waiting time (days)Adjusted median waiting time (days)a
I (high)118789.965
II (intermediate)312287.976
III (low)74286.577
  • aAdjusted for: age, gender, COPD, COPD current admission (CA), hypertension, other forms of chronic ischemic heart diseases, blood disorders, blood disorders CA, cerebrovascular diseases, cerebrovascular diseases CA, dementias, dementias CA, chronic renal diseases, chronic renal diseases CA, vascular disease.


In the last few years, there has been increased interest in evaluating the relationship between socioeconomic position and waiting times or outcomes after total hip replacement surgery [5, 14, 2224]. More socioeconomically disadvantaged people, in general, have experienced longer waiting times [22, 23] and worse outcomes [14, 24]. However, little is known about the effect of social factors on short-term mortality and time to surgery after hip fracture; surgery for acute hip fracture differs significantly from hip joint replacement as it is non-elective and the patient's needs and clinical course are more variable.

Several studies have demonstrated the advantages of early surgery in hip-fracture patients [6, 25, 26] and intervention within 24 h after hospital admission is recommended by the Royal College of Physicians' guidelines [27]. Recently, the Organization for Economic Co-operation and Development (OECD) included a 48-h waiting time to surgery for elderly patients with hip fracture in its national quality indicator list [28]. Also, a meta-analysis has shown delaying surgery, for 48 h or more after admission, may significantly increase the odds of 30-day and 1-year mortality [29]. However, few studies have evaluated the relationship between timing of hip-fracture surgery and the level of socioeconomic deprivation [36].

In our study, we found socioeconomic position was associated with both mortality and risk of early intervention. Individuals living in disadvantaged census tracts had a higher probability of death within 30 days of index or ED admission for hip fracture and were less likely to undergo surgery within 48 h of hospital arrival, when compared with more affluent people, even after we adjusted for available covariates. Lower socioeconomic level was also associated with decreased risk of intervention within 30 days. We also used hierarchical logistic regression to control for patient and hospital characteristics that could affect the outcomes under study. We found the classical and multilevel approaches gave similar results.

Overall, only 7.1% of interventions were performed within 48 h of index or ED admission. These findings are very different from those of other studies [25, 26, 30], although the range of performance of patients operated within 2 days, reported by western countries, varies from 33% to 93% [28]. Patients who underwent early intervention were less likely to have comorbidity variables. These findings confirmed comorbidities do indeed affect management, including the timing of operation. However, the differences between low and high socioeconomic position remained after adjustment for patients characteristics. According to other studies, men have a higher risk of death after hip fracture than women [7, 31]; yet, even after controlling for the effects of case mix, our results suggested worse social conditions enlarged discrepancies between genders.

This study has shown, for the first time in Italy, the association between socioeconomic position and quality of health care for hip fracture. Our results were obtained using the current healthcare information systems and, in particular, by the additional use of information from the HEIS to calculate time to death and time to surgery. In this respect, as increased time between hospital arrival and effective treatment for hip fracture may result in worse health outcomes, mortality and time to surgery were calculated from hospital arrival, corresponding to the date of index admission or ED admission. Moreover, we used the HEIS for additional information on patients' comorbidities, which were included in risk adjustment models.

The limits to this study should be considered. Time to surgery was computed based on the dates of hospital arrival and surgery and was not refined to the actual hour of surgery. Moreover, although a number of covariates were included in the models to account for differences in patient characteristics, unmeasurable differences among patients that affect the risk of 30-day mortality and intervention within 48 h may still remain. For example, the administrative data do not distinguish between acute conditions present at admission and complications of care. Therefore, we could not take into account those acute diseases (i.e. heart failure in the index admission) that are likely to affect risk of death and preoperative delay. As the need to stabilize concurrent medical conditions is a valid reason for delay, the longer waiting times experienced by the most disadvantaged may, to some degree, be related to their inferior baseline clinical conditions that could have resulted in delayed surgery. Another potential limitation is that by attributing to each patient an aggregated indicator of socioeconomic position, the true association could be underestimated. This has already been observed [32], even though the small area indicator is considered to be a good proxy of individual data [33]. On one hand, using a small area-based index of socioeconomic position, could capture characteristics of the area that are not captured using an individual index. On the other hand, the utilization of area-based indices as a proxy of individual traits could lead to a misclassification of individual socioeconomic position [34]. However, since the census tracts in Rome are rather small (average population: 500 inhabitants), the misclassification effect, if any, is likely to be low. As a consequence, our results may underestimate the differences relative to using direct measures of the socioeconomic position of individuals.

Moreover, it is possible that not all residents of a specific tract are served by the same hospital. However, in Rome, wealthier individuals generally tend to choose the hospital for an elective intervention but most of the residents in a specific neighborhood are transported by the ambulance and admitted to the nearest hospital for an emergency condition as acute hip fracture. It is interesting to highlight that in Rome acute care hospitals are not homogeneously distributed in the different city areas but no correlation seems to exist between more advantaged tracts and location of the best performing providers, identified by comparative outcome evaluations of Lazio hospitals [35].

In conclusion, this study has shown socioeconomic disparities in short-term mortality and time to surgery after hip fracture in Rome, further contributing to this complex issue. Our findings are in contrast to the principle of equity in the Italian Health Care System, where universal coverage should guarantee comparable quality of care to all citizens. Therefore, efforts by National and/or Regional Health Systems are needed to identify and address, in a systematic way, disparities that may be responsive to improvements in health care.


The authors would like to thank EUPHORIC (EUropean Public Health Outcome Research and Indicators Collection) project partners for their support.

Appendix: List of risk factors used for risk adjustment

Risk factorICD-9-CM code
Current admissionPrevious hospital or ED admissions
Diseases of central nervous system
 Dementias including Alzheimer's disease290.0–290.4, 294.1, 331.0290.0–290.4, 294.1, 331.0
 Parkinson's disease332332
 Hemiplegia and other paralytic syndromes342, 344342, 344
Ischemic heart disease
 Previous myocardial infarction412410, 412
 Other forms of chronic ischemic heart disease411, 413, 414
Other cardiac disorders
 Heart failure428
 Ill-defined descriptions and complications of heart disease429
 Rheumatic heart disease393–398391, 393–398
 Acute endocarditis and myocarditis421, 422
 Other heart conditions745, V15.1, V42.2, V43.2, V43.3, V45.0745, V15.1, V42.2, V43.2, V43.3, V45.0
 Cardiac arrhythmias426, 427
Circulatory disorders
 Cerebrovascular disease433, 437, 438430–434, 436–437, 438
 Vascular disease440–448 (excluding 441.1, 441.3, 441.5, 441.6, 444)440–448, 557
COPD491–492, 494, 496491–492, 494, 496
Chronic renal disease582–583, 585–588582–583, 585–588
Other chronic disease (liver, pancreas, intestine)571–572, 577.1–577.9, 555, 556571–572, 577.1–577.9, 555, 556
Bone and joint disorders
 Rheumatoid arthritis and other inflammatory polyarthropathies714714
 Osteoporosis and other disorders of bone and cartilage733733
Nutritional deficiencies260–263, 783.2, 799.4260–263, 783.2, 799.4
Blood disorders280–285, 288, 289280–285, 288, 289


View Abstract