General Description Over the Period of the Survey
Data from a total of 3972 patients (1081 men, 27.2%) and (2891 women, 72.8%) were available for analysis, with a mean age of 83.5 ± 9.1 years (median 84.9 years, IQR = 79.0–89.9), with a quarter of patients older than 90 years. Patients were analysed for all outcomes, except for the analysis of discharge destination where patients had to be from their own home originally and survived to discharge. Overall, there were 3077 patients (77.5%) who came from their own home, 24 (0.6%) from rehabilitation units, 529 (13.3%) from residential care, 246 (6.2%) from nursing care, 87 (2.2%) from hospital and 9 (0.2%) from other types of residence (Fig. 1).
The median elapsed time to surgery was 21.8 h (IQR = 16.8–29.8). Overall, 61.7%, 17.8% and 20.5% were operated within 24 h, 24–36 h and > 36 h, respectively (see also Table 1). Of these, 42.0% had ASA grade 3 and 5.9% with ASA grade 4, so that 48.0% had ASA grade ≥ 3 (including three with ASA grade 5; see Table 1). The median LOS during admission was 13.3 days (IQR = 8.3–23.3). The majority of patients received arthroplasty (51.9%), followed by similar proportions of treatment with an intramedullary nail (IMN) (22.4%) or a sliding hip screw (SHS) (24.4%); a small proportion received other techniques such as hybrid total hip replacement (1.3%). The majority of anaesthesia used spinal block with general anaesthetics (42.2%) or spinal block only (18.3%), whilst just over a quarter used general anaesthetics only.
Table 1 Outcomes of 3972 patients undergoing surgery for a hip fracture After hip surgery, 10% of patients failed to mobilise within one day, 3.3% developed a pressure ulcer, 5.7% died in hospital and 25.0% of patients stayed longer than 23 days in hospital. Amongst patients who were admitted from their own home, 59.2% returned back home, 25.1% were newly discharged to rehabilitation, 2.1% to residential care, 4.5% to nursing care and 4.0% to other residence. Antiresorptive therapy was used in only 4.6% of patients on admission and was prescribed in 83% of patients by discharge (Table 1). The majority of whom were prescribed oral anti-resorptive agents (93.5%), whilst only 6.5% received an injectable agent (denosumab). None of the patients were prescribed with anabolic agents such teriparatide or parathyroid hormone.
Trends in Perioperative Characteristics and Operative Outcomes
The proportions of patients operated beyond 36 h fell sharply during the first two years (2009–2011): − 54.1% (95% CI − 68.7, − 32.6, P = 0.003), followed by a small rise thereafter: APC = 6.3% (95% CI 0.9, 11.9, P = 0.029) (Fig. 2A, Table 2). During the first two years, the mean age at operation was lowest at about 82.5 years, rising gradually to a plateau of about 84 years four years later (Supplement Fig. 1). During the period of study, hip operation for patients > 90 years rose significantly (APC = 3.3%, 95% CI 1.0, 5.8, P = 0.011) (Fig. 2B) and those with ASA grade ≥ 3: APC = 12.4% (95% CI 8.8, 16.1, P < 0.001) (Fig. 2C). Over this time, the use of arthroplasty did not change, but there was an increased use of IMN: APC = 14.6% (95% CI 9.1, 20.3, P < 0.001) and a decrease of SHS: APC = − 13.5% (95% CI − 18.1, − 6.2, P < 0.001) (Supplement Fig. 2). There were no changes in patient numbers coming from nursing care.
Table 2 Annual percentage change for amongst patients undergoing operation for hip fractures Trends in Post-operative Outcomes
There were no significant changes in failure to mobilise within one day of hip surgery. Table 2 shows that although there was an overall decline in the rates of pressure ulcers and mortality, this did not reach a significance level (Fig. 3A and B). However, when analysed separately by age groups (by upper quartile threshold: 90 years), there was a significant decline in pressure ulcers amongst patients under 90 years old: APC = − 17.9 (95% CI − 32.7, 0.0, P = 0.050) and also a significant decline in mortality amongst those over 90 years old: APC = − 7.1 (95% CI − 12.6, − 1.3, P = 0.024) (Fig. 3C and D).
The median LOS in hospital in the first two years (2009–2011) was just over 15 days which reduced relatively little until a high reduction from 2013 to 2019 (Fig. 4A). There was a single join point at 2013 that coincided with data collection of the new service (see Introduction) (Fig. 4B); the APC did not change between 2009 and 2013, whilst there was a significant decline from 2013 to the end of the study period (APC = − 24.6, 95% CI − 31.2, − 17.4, P < 0.001) (Table 2). When no join point was used, the decline of the AAPC (= − 15.2, 95% CI − 21.1, − 8.8, P = 0.001) remained significant (Fig. 4C).
Trends in Discharge
There was a single join point at 2016, coinciding with introduction of the COW model (see Introduction) (Fig. 5A); there was a moderate APC between 2009 and 2016 (APC = − 10.6, 95% CI − 17.2, − 2.7, P = 0.017), followed by a sharper decline between 2016 and the end of the study period (APC = − 47.5, 95% CI − 71.7, − 2.7, P = 0.043). The AAPC was − 17.5 (95% CI − 24.3, − 10.1, P = 0.001) (Fig. 5B).
There was a continuously decreasing trend in the proportions of patients returning home (APC = − 2.9, 95% CI − 5.1, − 0.7, P = 0.016) (Fig. 5C), whilst the trend in new discharge to rehabilitation increased (APC = 8.4, 95% CI 4.0, 13.0; P = 0.002) (Fig. 5D).
The proportion of patients newly prescribed with an anti-resorptive agent was only 61% in 2010 when two orthogeriatricians were appointed later in that year (September 2010), rising significantly to 85.3% since 2011 (group difference: χ2 = 159, P < 0.001) (Fig. 6).