Health care consumption and costs due to foot and ankle injuries in the Netherlands, 1986–2010

Data sources

For this retrospective study data were collected for patients with foot and ankle injuries in the Netherlands in the period 1986–2010. Injury cases were extracted from the National Injury Surveillance System (LIS) [27] and National Medical Registration (LMR) [28], to include non-hospitalized and hospitalized patients, respectively. LIS is a continuous monitoring system that records unintentional and intentional injuries. It has been implemented in 17 hospital EDs, resulting in a representative 12% sample of all injury-related ED visits in the Netherlands [27]. These hospitals are geographically distributed across the country with their adherence population being representative for the Dutch population in age and gender structure [29]. LMR collects data regarding hospital admissions, admission diagnosis, gender, age, and length of hospital stay. LMR is centrally evaluated for plausibility and completeness before entry into the LIS database [27]. LMR has almost complete national coverage (<5% missing except 12% for 2007) and figures are extrapolated to full national coverage for each year. An extrapolation factor was determined by comparing the adherence population of the participating hospitals with the total Dutch population in each year [28]. Patients are included in LIS and LMR according to their main diagnosis at discharge, which is generally the most severe injury. Coding of patients was consistently based upon full patient chart review including routine radiological assessment as available in the patient files.

Data regarding hospital length of stay (HLOS) were extracted from the LMR database for 10-year age categories. In order to assess trends in HLOS over time, the mean HLOS was averaged over 5-year intervals from 1991–2010. The time periods for the different analyses (1986–2010 for incidence rates, 1991–2010 for HLOS, and 2010 for health care consumption and associated costs) were based on data availability.

The study was exempted by the local Medical Research Ethics Committee Erasmus MC (No. MEC-2014-006).

Calculation of emergency attendance rates and incidence rates for hospital admission

Data were analyzed using the Statistical Package for the Social Sciences (SPSS) version 16.0 for Windows.

Age-specific emergence attendance rates (for all patients presented to the ED) and incidence rates for hospital admission (for all patients admitted to hospital) were calculated in 5-year age groups. This was done for the total population and for males and females separately. For each age group the absolute numbers of hospitalized and non-hospitalized cases with foot and ankle injuries were extracted from the LMR and LIS database, respectively. Since patient numbers in the LIS database were obtained from a sample, they were weighted in order to create national estimates. An extrapolation factor was determined by comparing the number of admitted injury patients in the LIS database with the total number of admitted injury patients in the LMR database. In order to adjust for differences in the demographic composition over time, emergency attendance rates and incidence rates for hospital admission were standardized for age (in 5-year age groups) and gender using a direct standardization method. The age- and gender-specific emergency attendance rates and incidence rates for hospital admission per 100,000 person years were calculated based upon the Dutch mid-year standard population. Mid-year population sizes for all age groups were obtained from Statistics Netherlands [31]. Age-adjusted emergency attendance rates and incidence rates for hospital admission were calculated using “direct standardization” [32]. The average number of persons in each 5-year age class for each year of the study (1986–2010) was calculated. This number was used as the standard (reference) population, as described previously [33, 34]. Overall increase in hospital admissions was calculated for 2010 in per cents relative to 1986.

Calculation of costs

The incidence-based Dutch Burden of Injury Model, which has been used in ten European countries, was used in order to measure and describe the health care costs for the year 2010 [24, 33, 35–38]. Patient numbers, health care consumption, and related costs were calculated for the four injury groups using the LIS database, the National Hospital Discharge Registry, and a patient follow-up survey to calculate associated direct health care costs in 2010. The patient follow-up survey collected data on in-hospital care, outpatient visits, general practitioner (G.P.) visits, outpatient physical therapy, home care, medication, and aids and appliances [29]. Costs and health care consumption are injury-, gender- and age-dependent. In our model, the age- and injury-specific costs were based upon the estimated health care supplied to the individual patients. Costs were determined for the following categories: 1) ambulance care; 2) G.P. visits; 3) in-hospital care; 4) home care; 5) rehabilitation and nursing home care; and 6) physical therapy. Health care costs of injuries were calculated by multiplying incidence, health care volumes (e.g., length of stay in hospital or institution, the number of outpatient visits, G.P. visits, home care hours, and physical therapy treatments) with unit costs (e.g., costs per day in hospital). Unit costs were estimated according to national guidelines for health care costing [38]. Age-specific costs are presented in 10-year age groups for men and women separately.

Emergency attendances and hospital admissions

During the study period, the absolute number of patients reporting to an ED with a foot or ankle injury decreased from 124,595 in 1986 to 106,157 in 2010. The emergency attendance rate of all injuries combined decreased from 858 to 640 per 100,000 persons (-25.4%). Whereas ligamentous injuries approximately halved, osseous injuries nearly doubled.

In non-admitted patients, representing 90% of patients, the overall emergency attendance rate decreased by 30.2% (Figure 1A). This was mainly due to a decrease in ligamentous injuries of the ankle (504/100,000 in 1986 versus 228/100,000 in 2010; -54.8%) and foot (26/100,000 in 2010; -50.9%). Osseous injuries in the foot and ankle, however, increased by 28.3% (152/100,000 in 2010) and 25.3% (104/100,000 in 2010), respectively.

The admission rate increased from 7.6% in 1986 to 13.8% in 2010 (Figure 1B). This was mainly due to a 31.8% admission rate of patients with osseous ankle injuries. Admission of patients with osseous foot injuries (4.0% admitted) or ligamentous injuries in the ankle (3.8%) or foot (<0.1%) was low. Since 1986, the incidence rate for hospital admission due to foot and ankle injuries increased by 35.4%. This was mostly due to increased incidences of osseous ankle injuries (33/100,000 in 1986 versus 62/100,000 in 2010; +87.9%). The incidence rate for hospital admission due to ligamentous ankle injuries diminished with 42.3% (26 to 15/100,000).

The emergency attendance rates of foot and ankle injuries varied with age in males and, less pronounced, also in females (Figure 2A and 2B). Attendance rates showed a peak in adolescents and adults until ~45 years of age. Until this age the attendance rate in males was higher than in females. Since 1986 this peak in attendance has decreased in both genders. The decrease in incidence peaks at younger ages over time suggests a shift towards a higher mean age. Indeed, the mean age of patients increased throughout the study period by 8.3 years for osseous ankle injuries (from 32.1 ± 19.7 (SD) years in 1986 to 40.4 ± 22.7 in 2010) and by 4.0 years for osseous foot injuries (from 31.6 ± 18.5 in 1986 to 35.5 ± 20.9 in 2010), Mean ages were much more stable for ligamentous injuries (age increased from 25.4 ± 15.7 to 26.7 ± 17.9 years for the foot and from 27.0 ± 13.4 to 29.6 ± 16.7 years for the ankle).

Figure 2C and 2D show age-trends of the four main injury types in 2010. Again, a peak in adolescents was seen, especially in males. Whereas the incidence in all injury types reduced with age in males, the incidence of osseous injuries in elderly women remained more stable.

Hospital length of stay

Hospital length of stay (HLOS) in four consecutive five-year periods is shown in Figure 3A and 3B. Each period showed a gradual increase with age, yet over time the HLOS decreased for all age groups. The HLOS per case more than halved both in males (7.8 days in 1991 versus 3.3 in 2010) and females (11.5 days versus 4.9). The total number of hospital days for men and women of all ages combined decreased from 78,951 days in 1991 to 58,708 days in 2010. Patients aged 20–65 year accounted for 51% of all hospital days.

The HLOS for different types of injuries is shown in Figure 3C and 3D (for males and females, respectively). Osseous injuries caused the longest hospital stay per case in almost every age group, with limited differences between foot and ankle injuries. HLOS in males and females was similar for all injury types; in every age group the difference was restricted to one day at most. Exceptions to this were noted in elderly (70+) with an ankle injury; HLOS for osseous ankle injuries was 11.0 days in females versus 8.2 days in males. HLOS for ligamentous ankle injuries was 5.3 and 3.4 days, respectively.

Costs for health care consumption

The overall costs for all patients amounted 161.9 million euro in 2010. Between 2001 and 2010, costs remained fairly stable; an overall increase of 1.2% was noted (data not shown).

Figure 4 shows the costs per case for the four main injury groups, separated into costs for different types of health care use. In addition to costs per case being higher in females than in males, costs consistently increased with age for all four injury categories; from 0 to 70+ years, costs for all injuries combined increased 9.2-fold in females and 5.7-fold in males. The largest increase with age was seen for osseous injuries of the ankle (from € 996 to € 6,023 in males and from € 1,127 to € 10,949 in females; Figures 4E and F) and foot (from € 571 to € 1,716 in males and from € 642 to € 3,626 in females; Figures 4A and B). Ligamentous foot injuries showed only a 1.7-fold and 2.0-fold increase across the age groups in males and females, respectively (Figures 4C and D). Costs for ligamentous injuries were independent of gender. For osseous injuries, the costs per case were similar in males and females until 60 years of age, but a clear gender-dependency was noted for the 70+ group.

Costs for in-hospital care consistently contributed most to the total cost per case (64-69% of total costs in males and 45-62% in females; Figure 4). Physical therapy was the second largest determinant (8-19% of total costs). Ambulance care and G.P. visits each contributed less than 10% to the overall cost per case; they doubled or tripled over the age groups, but were unrelated to injury type and gender. For osseous foot and ankle injuries, the increase in costs over the age groups was mainly due to increased use of rehabilitation/nursing care and home care in the elderly. This effect was more pronounced in females than in males. The age-effect on costs in ligamentous foot injuries was mainly attributable to increased costs for physical therapy, G.P. visits, and home care. Higher use of home care also explained the larger cost increase in females. For ligamentous ankle injuries, in-hospital costs increased more with age in males, whereas costs for home care and physical therapy increased more in females.