Browse the corpus

SNAP-3: characteristics of specific subgroups of older patients undergoing surgery in the UK. BACKGROUND: The SNAP-3 study reported one in five older surgical patients in the UK were living with frailty and two in three with multimorbidity. We now report characteristics and outcomes of specific subgroups of patients including individuals aged ≥85 yr, undergoing day or inpatient surgery and elective or emergency surgery, and undergoing common specialty procedures including orthopaedics, urology, colorectal surgery, and hip arthroplasty. METHODS: This prospective observational cohort study recruited patients aged ≥60 yr undergoing surgery during five days in March 2022 across 214 UK hospitals. Daycase, inpatient, elective, and emergency surgery was included; minor or superficial surgery was excluded. Data on patient characteristics, frailty, and multimorbidity were collected. Outcomes were medical complications including delirium, collected prospectively, postoperative morbidity and death, and length of stay. Descriptive statistics were used to analyse demographic and outcome data for these subgroups defined a priori. RESULTS: Of 7821 participants recruited, 7134 were available for analysis. Frailty was most prevalent amongst those undergoing hip fracture surgery (58.2%; 280/481); aged ≥85 yr (49.3%; 365/740); or undergoing emergency (31.4%; 668/2126) or orthopaedic procedures (30.6%; 630/2056). Participants undergoing emergency procedures had a median postoperative stay of 5 (interquartile range 1-13) days versus 1 (interquartile range 0-3) days compared with participants undergoing elective procedures, with delirium occurring in 16.3% (348/2138) versus 2.7% (134/4922), respectively. Postoperative morbidity and 30-day mortality were highest in those undergoing hip fracture surgery (morbidity 85.3% [413/484], mortality 5.0% [24/484]); aged ≥85 yr (morbidity 49.1% [367/747], mortality 3.9% [29/741]) and undergoing emergency procedures (morbidity 49.5% [1058/2138], and mortality 3.0% [64/2124]). CONCLUSIONS: This snapshot of surgical patients in the UK highlights the high prevalence of frailty, postoperative morbidity, and mortality in key surgical subgroups, reinforcing the urgent need for comprehensive age-appropriate perioperative services tailored to these populations.

This study was undertaken to assess the clinical characteristics of subgroups of interest prespecified by the SNAP-3 study management group as described.12 In summary, all UK hospitals that deliver adult surgical services were invited to participate in a prospective observational cohort study. The study aimed to recruit all patients aged ≥60 yr, undergoing a surgical procedure during five consecutive days (Monday–Friday) in March 2022. The study was conducted between waves of the COVID-19 pandemic, and despite the challenges this posed, aimed to recruit a representative cohort. A comparison of SNAP-3 data with Hospital Episode Statistics and National Audit Project data suggests that the cohort is representative of the broader older surgical population in the UK.13,14 The inclusion criteria were deliberately broad: those aged ≥60 yr undergoing almost any type of surgery under general, regional, neuraxial, or local anaesthesia. Only very minor procedures such as cataract surgery or diagnostic endoscopy were excluded (Supplementary Table S1). Local sites were encouraged to include those without capacity to consent to the study using a consultee or personal legal representative. Ethical approval was provided by the Wales Research Ethics Service (21/WA/0203) and the Scotland A Research Ethics Committee (302033) in July and September 2021, respectively. Electronically recorded or written informed consent/assent was obtained for all participants.

sing a consultee or personal legal representative. Ethical approval was provided by the Wales Research Ethics Service (21/WA/0203) and the Scotland A Research Ethics Committee (302033) in July and September 2021, respectively. Electronically recorded or written informed consent/assent was obtained for all participants. Subgroups included those undergoing day vs inpatient surgery; elective vs non elective procedures (as defined by National Confidential Enquiry into Patient Outcome and Death [NCEPOD] criteria15); operations within the three most prevalent surgical specialties; those aged ≥85 yr vs those <85 yr; and individuals undergoing hip fracture surgery. In the UK, day surgery is defined as surgery conducted where the patient is admitted and discharged on the same day.9 Covariates included demographic data; medical and surgical history; laboratory data; SARS-CoV-2 status; surgical risk scores; and socioeconomic data collected by local investigators. Characteristics known to be associated with adverse outcomes in older people were collected including: frailty; multimorbidity; participant’s residence; polypharmacy; independence in activities of daily living (ADLs); and dementia. Frailty status was determined by the Clinical Frailty Score (CFS, dichotomised as frail ≥CFS 5).16 Multimorbidity was reported as a count of two or fewer specified comorbidities (from a list adapted from the Charlson Comorbidity Index17 to include other conditions known to be relevant in the perioperative period, for example obstructive sleep apnoea and atrial fibrillation, Supplementary Table S2).

frail ≥CFS 5).16 Multimorbidity was reported as a count of two or fewer specified comorbidities (from a list adapted from the Charlson Comorbidity Index17 to include other conditions known to be relevant in the perioperative period, for example obstructive sleep apnoea and atrial fibrillation, Supplementary Table S2). Outcomes for the SNAP-3 study overall included medical complications (delirium, postoperative morbidity, death); healthcare resource used (length of stay, days alive and out of hospital, readmissions); and quality of life. This report focuses on demographic data, frailty, multimorbidity, and the outcomes of length of stay, delirium and other in-hospital morbidity, and mortality to 1 yr. Participants who remained inpatients on days 3 and 7 were assessed for postoperative morbidity using an appropriate specialty specific PostOperative Morbidity Survey18, 19, 20 (POMS) and either the 4-As test (4AT),21 if not critically ill, or Confusion Assessment Method for the ICU (CAM-ICU),22 if critically ill. Delirium and postoperative morbidity were assumed absent for those discharged alive on the day of surgery unless documented.

cialty specific PostOperative Morbidity Survey18, 19, 20 (POMS) and either the 4-As test (4AT),21 if not critically ill, or Confusion Assessment Method for the ICU (CAM-ICU),22 if critically ill. Delirium and postoperative morbidity were assumed absent for those discharged alive on the day of surgery unless documented. Those admitted for one or more nights underwent a retrospective notes review to identify delirium with the aim of minimising false negatives from researcher assessments alone given the fluctuating nature of the condition. Similarly, POMS was assumed to be absent if patients were discharged before day 3 or 7. This included medical and nursing documentation from the day of surgery, up to discharge or day 7 after surgery, whichever was sooner. Population-based healthcare administration records (NHS Digital, Digital Health and Care Wales, and NHS National Services Scotland) were used to collate further information regarding readmission, discharge, and mortality. Local investigators were resident anaesthetists and physicians, and research nurses, supported by consultant anaesthetists. All local investigators were directed to web-based training in the completion of frailty scores.23,24 We report estimated proportions of key demographic and preadmission data considered to be important predictors of outcomes in older people, and of the outcomes of length of hospital stay, delirium, postoperative morbidity, and mortality.

Those admitted for one or more nights underwent a retrospective notes review to identify delirium with the aim of minimising false negatives from researcher assessments alone given the fluctuating nature of the condition. Similarly, POMS was assumed to be absent if patients were discharged before day 3 or 7. This included medical and nursing documentation from the day of surgery, up to discharge or day 7 after surgery, whichever was sooner. Population-based healthcare administration records (NHS Digital, Digital Health and Care Wales, and NHS National Services Scotland) were used to collate further information regarding readmission, discharge, and mortality. Local investigators were resident anaesthetists and physicians, and research nurses, supported by consultant anaesthetists. All local investigators were directed to web-based training in the completion of frailty scores.23,24 We report estimated proportions of key demographic and preadmission data considered to be important predictors of outcomes in older people, and of the outcomes of length of hospital stay, delirium, postoperative morbidity, and mortality. Statistical analysis is purely descriptive as our objective was to report the characteristics and outcomes of these clinically relevant subgroups. Detailed modelling of association with outcomes (as was performed for the main analyses12) was not undertaken. Data are presented for complete data (without imputation) and are presented as proportions, mean (sd), and median (IQR) as appropriate. Confidence intervals (95%) of proportions were obtained using bootstrapping. All analyses were conducted in R (version 4.3.1, R Project for Statistical Computing, Vienna, Austria).25 This manuscript is reported in accordance with STROBE guidelines (Supplementary Table S3).

The inclusion criteria were deliberately broad: those aged ≥60 yr undergoing almost any type of surgery under general, regional, neuraxial, or local anaesthesia. Only very minor procedures such as cataract surgery or diagnostic endoscopy were excluded (Supplementary Table S1). Local sites were encouraged to include those without capacity to consent to the study using a consultee or personal legal representative. Ethical approval was provided by the Wales Research Ethics Service (21/WA/0203) and the Scotland A Research Ethics Committee (302033) in July and September 2021, respectively. Electronically recorded or written informed consent/assent was obtained for all participants.

Subgroups included those undergoing day vs inpatient surgery; elective vs non elective procedures (as defined by National Confidential Enquiry into Patient Outcome and Death [NCEPOD] criteria15); operations within the three most prevalent surgical specialties; those aged ≥85 yr vs those <85 yr; and individuals undergoing hip fracture surgery. In the UK, day surgery is defined as surgery conducted where the patient is admitted and discharged on the same day.9

Covariates included demographic data; medical and surgical history; laboratory data; SARS-CoV-2 status; surgical risk scores; and socioeconomic data collected by local investigators. Characteristics known to be associated with adverse outcomes in older people were collected including: frailty; multimorbidity; participant’s residence; polypharmacy; independence in activities of daily living (ADLs); and dementia. Frailty status was determined by the Clinical Frailty Score (CFS, dichotomised as frail ≥CFS 5).16 Multimorbidity was reported as a count of two or fewer specified comorbidities (from a list adapted from the Charlson Comorbidity Index17 to include other conditions known to be relevant in the perioperative period, for example obstructive sleep apnoea and atrial fibrillation, Supplementary Table S2).

Outcomes for the SNAP-3 study overall included medical complications (delirium, postoperative morbidity, death); healthcare resource used (length of stay, days alive and out of hospital, readmissions); and quality of life. This report focuses on demographic data, frailty, multimorbidity, and the outcomes of length of stay, delirium and other in-hospital morbidity, and mortality to 1 yr. Participants who remained inpatients on days 3 and 7 were assessed for postoperative morbidity using an appropriate specialty specific PostOperative Morbidity Survey18, 19, 20 (POMS) and either the 4-As test (4AT),21 if not critically ill, or Confusion Assessment Method for the ICU (CAM-ICU),22 if critically ill. Delirium and postoperative morbidity were assumed absent for those discharged alive on the day of surgery unless documented.

cialty specific PostOperative Morbidity Survey18, 19, 20 (POMS) and either the 4-As test (4AT),21 if not critically ill, or Confusion Assessment Method for the ICU (CAM-ICU),22 if critically ill. Delirium and postoperative morbidity were assumed absent for those discharged alive on the day of surgery unless documented. Those admitted for one or more nights underwent a retrospective notes review to identify delirium with the aim of minimising false negatives from researcher assessments alone given the fluctuating nature of the condition. Similarly, POMS was assumed to be absent if patients were discharged before day 3 or 7. This included medical and nursing documentation from the day of surgery, up to discharge or day 7 after surgery, whichever was sooner. Population-based healthcare administration records (NHS Digital, Digital Health and Care Wales, and NHS National Services Scotland) were used to collate further information regarding readmission, discharge, and mortality. Local investigators were resident anaesthetists and physicians, and research nurses, supported by consultant anaesthetists. All local investigators were directed to web-based training in the completion of frailty scores.23,24

We report estimated proportions of key demographic and preadmission data considered to be important predictors of outcomes in older people, and of the outcomes of length of hospital stay, delirium, postoperative morbidity, and mortality. Statistical analysis is purely descriptive as our objective was to report the characteristics and outcomes of these clinically relevant subgroups. Detailed modelling of association with outcomes (as was performed for the main analyses12) was not undertaken. Data are presented for complete data (without imputation) and are presented as proportions, mean (sd), and median (IQR) as appropriate. Confidence intervals (95%) of proportions were obtained using bootstrapping. All analyses were conducted in R (version 4.3.1, R Project for Statistical Computing, Vienna, Austria).25 This manuscript is reported in accordance with STROBE guidelines (Supplementary Table S3).

Of the 263 NHS hospitals across the UK invited to participate, 214 participated, recruiting 7821 patients over a 5-day period. Of these, 687 patients were withdrawn from the study leaving data from 7134 participants for analysis.

Of the 263 NHS hospitals across the UK invited to participate, 214 participated, recruiting 7821 patients over a 5-day period. Of these, 687 patients were withdrawn from the study leaving data from 7134 participants for analysis. Key clinical characteristics of the participants in the subgroups are summarised in Table 1 (participants aged ≥85 yr vs younger participants; daycase vs inpatient surgery; and elective vs non elective surgery) and Table 2 (orthopaedic, urology, colorectal, and hip fracture surgery). More details of all characteristics are provided in Supplementary Tables S4, S5, S6 and S7. Further stratification of the subgroup characteristics according to frail/not frail and elective/non elective surgery are provided in Supplementary Tables S8 and S9. Postoperative outcomes are reported in Table 3 (participants aged ≥85 yr vs younger participants; daycase vs inpatient surgery; and elective vs non elective surgery) and Table 4 (orthopaedic, urology, colorectal, and hip fracture surgery). Further stratification of the subgroup outcomes according to frail/not frail and elective/non elective surgery are provided in Supplementary Tables S10 and S11. Data on mode of anaesthesia are tabulated for reference in Supplementary Tables S12–S15. Because of missingness, not all denominators are the same. Data on missingness are reported in Supplementary Table S16.Table 1Participant characteristics of the subgroups: 85 yr and older, 60–84 yr, daycase surgery, inpatient surgery, elective, and non elective surgery. The participant characteristics of the SNAP-3 subgroups. Percentages have been rounded so may not total 100% exactly. Missing data are omitted from this table and reported in Supplementary Table S9. Surgical urgency is defined using NCEPOD categorisations.15 The values here are proportions (%) with 95% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps.

exactly. Missing data are omitted from this table and reported in Supplementary Table S9. Surgical urgency is defined using NCEPOD categorisations.15 The values here are proportions (%) with 95% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps. Surgical urgency is described using the NCEPOD criteria.15 ASA, American Society of Anesthesiologists; NA, not applicable; NCEPOD, National Confidential Enquiry into Patient Outcome and Death.Table 1Characteristic≥85 Yr<85 YrDaycaseInpatientsElectiveNon-electiveSummaryNumberSummaryNumberSummaryNumberSummaryNumberSummaryNumberSummaryNumberAge (yr)Mean88 (87.8–88.3)74771 (70.9–71.2)630971 (70.7–71.3)240673.8 (73.6–74)465071.9 (71.7–72.1)492075 (74.6–75.4)213560–69NANA41 (39.8–42.1)2585/630944.8 (42.8–46.7)1079/240632.4 (31–33.8)1506/465039 (37.6–40.4)1920/492031.1 (29.1–33.2)665/213570–79NANA45 (43.7–46.3)2839/630939.2 (37.2–41.1)942/240640.8 (39.4–42.2)1897/465042.4 (41–43.8)2084/492035.3 (33.3–37.4)754/213580–8969.1 (65.9–72.4)516/74714 (13.2–14.9)885/630915 (13.6–16.5)361/240622.4 (21.2–23.6)1040/465017.3 (16.2–18.4)853/492025.7 (23.8–27.6)548/2135>9030.9 (27.4–34.3)231/747NANA1 (0.6–1.4)24/24064.5 (3.8–5.1)207/46501.3 (1–1.6)63/49207.9 (6.7–9.1)168/2135ASA physical status10.9 (0.3–1.8)7/7398.1 (7.4–8.8)506/625511.3 (10.1–12.6)270/23825.3 (4.6–5.9)243/46178.1 (7.3–8.8)395/48725.5 (4.6–6.5)118/2127231.3 (28–34.5)231/73955.8 (54.5–57.1)3491/625563.6 (61.8–65.6)1516/238247.8 (46.4–49.3)2209/461758.6 (57.2–60.1)2856/487240.9 (38.7–42.9)869/2127357 (53.3–60.6)421/73933.1 (31.9–34.2)2068/625523.9 (22.2–25.6)569/238241.6 (40.2–43)1921/461731.5 (30.2–32.8)1533/487245 (42.9–47.1)957/2127410.4 (8.3–12.6)77/7392.9 (2.5–3.4)183/62551 (0.7–1.5)25/23825.1 (4.5–5.8)236/46171.7 (1.3–2)82/48728.4 (7.3–9.6)179/212750.4 (0–0.9)3/7390.1 (0–0.2)7/62550.1 (0–0.2)2/23820.2 (0.1–0.3)8/46170.1 (0–0.2)6/48720.2 (0–0.4)4/2127Frailty (CFS ≥5)49.3 (45.8–52.8)365/74016 (15.2–17)1003/625210.6 (9.4–11.9)253/238024.2 (22.9–25.4)1116/461714.4 (13.4–15.4)701/487031.4 (29.4–33.3)668/2126Clinical Frailty Scale12.2 (1.2–3.4)16/74012.7 (11.8–13.5)792/625216.3 (15–17.9)389/23809.1 (8.3–9.9)419/461712.4 (11.5–13.3)603/48709.6 (8.4–10.9)205/212627 (5.3–8.9)52/74020.8 (19.8–21.8)1298/625225.7 (23.9–27.5)612/238016 (14.9–17.1)739/461720.7 (19.5–21.8)1008/487016.1 (14.5–17.7)343/2126317.3 (14.6–20)128/74032.2 (31.1–33.3)2012/625233.5 (31.6–35.3)797/238029.1 (27.8–30.4)1344/461733 (31.7–34.4)1609/487025 (2

9.9)419/461712.4 (11.5–13.3)603/48709.6 (8.4–10.9)205/212627 (5.3–8.9)52/74020.8 (19.8–21.8)1298/625225.7 (23.9–27.5)612/238016 (14.9–17.1)739/461720.7 (19.5–21.8)1008/487016.1 (14.5–17.7)343/2126317.3 (14.6–20)128/74032.2 (31.1–33.3)2012/625233.5 (31.6–35.3)797/238029.1 (27.8–30.4)1344/461733 (31.7–34.4)1609/487025 (2 3.1–26.9)531/2126424.2 (21.1–27.3)179/74018.3 (17.4–19.3)1147/625213.8 (12.4–15.2)329/238021.6 (20.4–22.8)999/461719.5 (18.4–20.6)949/487017.8 (16.2–19.4)379/2126522.3 (19.2–25.4)165/7408.5 (7.8–9.1)530/62526 (5–7)143/238012 (11.1–12.9)553/46178.6 (7.8–9.3)418/487013.1 (11.7–14.5)278/2126614.5 (11.9–17)107/7405 (4.5–5.6)314/62523.6 (2.8–4.3)85/23807.3 (6.5–8)336/46174.3 (3.8–4.9)210/48709.9 (8.7–11.2)211/2126710.8 (8.6–13)80/7402.3 (1.9–2.7)142/62521 (0.6–1.4)24/23804.3 (3.7–4.9)198/46171.4 (1–1.7)67/48707.3 (6.2–8.4)155/212681.5 (0.7–2.4)11/7400.2 (0.1–0.4)14/6252NANA0.5 (0.3–0.8)25/46170.1 (0–0.2)5/48700.9 (0.6–1.4)20/212690.3 (0–0.7)2/7400 (0–0.1)3/62520 (0–0.1)1/23800.1 (0–0.2)4/46170 (0–0.1)1/48700.2 (0–0.4)4/2126Multimorbidity (≥2 comorbidities)80.6 (77.7–83.6)561/69660.9 (59.6–62.3)3415/560455.1 (52.9–57.2)1089/197766.8 (65.3–68.2)2889/432661.1 (59.6–62.6)2650/433967.6 (65.5–69.7)1328/1964Surgical urgencyEmergency4.7 (3.2–6.3)35/7472 (1.6–2.3)125/63080.5 (0.2–0.8)12/24063.2 (2.7–3.7)149/4654NANA7.5 (6.5–8.7)161/2138Urgent35.6 (32.3–39)266/74712.6 (11.8–13.5)795/63084.9 (4.1–5.8)119/240620.3 (19.2–21.4)943/4654NANA49.7 (47.6–51.8)1062/2138Expedited14.7 (12.2–17.4)110/74712.7 (11.9–13.6)804/630810.1 (9–11.3)243/240614.4 (13.5–15.4)672/4654NANA42.8 (40.7–44.9)915/2138Planned45 (41.4–48.6)336/74772.7 (71.6–73.7)4584/630884.5 (82.9–85.9)2032/240662.1 (60.7–63.5)2890/4654100 (NA–NA)4922/4922NANA Daycase18.5 (15.7–21.3)138/74735.9 (34.8–37.2)2268/6309100 (NA–NA)2407/2407NANA41.3 (39.9–42.6)2032/492217.5 (16–19.1)374/2138Table 2Participant characteristics of the subgroups: orthopaedic, colorectal, urological, and hip fracture surgery.

630884.5 (82.9–85.9)2032/240662.1 (60.7–63.5)2890/4654100 (NA–NA)4922/4922NANA Daycase18.5 (15.7–21.3)138/74735.9 (34.8–37.2)2268/6309100 (NA–NA)2407/2407NANA41.3 (39.9–42.6)2032/492217.5 (16–19.1)374/2138Table 2Participant characteristics of the subgroups: orthopaedic, colorectal, urological, and hip fracture surgery. The participant characteristics of the SNAP-3 subgroups. Percentages have been rounded so may not total 100% exactly. Missing data are omitted from this table and reported in Supplementary Table S9. Surgical urgency is defined using NCEPOD categorisations.15 The values here are proportions (%) with 95% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps.

exactly. Missing data are omitted from this table and reported in Supplementary Table S9. Surgical urgency is defined using NCEPOD categorisations.15 The values here are proportions (%) with 95% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps. Surgical urgency is described using the NCEPOD criteria.15 ASA, American Society of Anesthesiologists; NA, not applicable; NCEPOD, National Confidential Enquiry into Patient Outcome and Death.Table 2CharacteristicOrthopaedicsColorectalUrologyHip fractureSummaryNumberSummaryNumberSummaryNumberSummaryNumberAge (yr)Mean74.5 (74.1–74.8)207073 (72.5–73.6)85073.6 (73.2–74)133681.6 (80.9–82.5)48360–697.1 (6–8.2)146/20702 (1.1–2.9)17/8502.4 (1.6–3.2)32/133621.3 (17.6–25.3)103/48370–7932 (30–34)662/207036.5 (33.2–39.8)310/85031.6 (29.1–34.1)422/133611.8 (8.9–14.7)57/48380–8937.5 (35.5–39.6)777/207038.1 (34.7–41.5)324/85042.4 (39.7–45.1)566/133623.6 (19.9–27.5)114/483>9023.4 (21.6–25.4)485/207023.4 (20.7–26.2)199/85023.7 (21.3–26)316/133643.3 (38.9–47.8)209/483ASA physical status17.4 (6.2–8.5)152/20597.7 (5.9–9.5)65/8436.3 (5–7.5)83/13262.1 (0.8–3.3)10/483250.1 (48.1–52.3)1032/205952.7 (49.5–56)444/84357.8 (55.1–60.6)767/132627.5 (23.6–31.5)133/483337.2 (35–39.3)765/205936.7 (33.5–40)309/84333.9 (31.4–36.4)450/132655.9 (51.3–60.2)270/48345.2 (4.3–6.2)107/20592.7 (1.7–3.9)23/8432 (1.2–2.7)26/132614.3 (11.4–17.6)69/48350.1 (0–0.3)3/20590.2 (0–0.6)2/843NANA0.2 (0–0.6)1/483Frailty (CFS ≥5)30.6 (28.6–32.7)630/205613.2 (10.9–15.4)111/84315.4 (13.5–17.5)204/132358.2 (53.8–62.8)280/481Clinical Frailty Scale18 (6.9–9.3)165/205613.9 (11.5–16.3)117/84312.6 (10.8–14.5)167/13233.3 (1.9–5)16/481214.1 (12.5–15.7)290/205621.4 (18.6–24.1)180/84321.8 (19.6–24)288/13236 (4–8.3)29/481325 (23.2–26.9)514/205633.9 (30.7–37.2)286/84334.4 (31.9–37)455/132317.3 (13.9–20.8)83/481422.2 (20.3–24.1)457/205617.7 (15.1–20.3)149/84315.8 (13.8–17.8)209/132315.2 (12.1–18.5)73/481514.3 (12.7–15.8)293/20568.3 (6.5–10.2)70/8438.5 (7–10)112/132322 (18.3–25.8)106/48169.2 (8–10.6)190/20563.3 (2.1–4.5)28/8434.4 (3.3–5.5)58/132317.3 (13.9–20.8)83/48176.3 (5.3–7.4)129/20561.2 (0.5–2)10/8432.2 (1.5–3)29/132316.4 (13.3–19.8)79/48180.8 (0.4–1.2)16/20560.4 (0–0.8)3/8430.2 (0–0.4)2/13232.3 (1–3.7)11/48190.1 (0–0.2)2/2056NANA0.2 (0–0.5)3/13230.2 (0–0.6)1/481Multimorbidity(≥2 comorbidities)66.2 (63.9–68.3)1216/183757.7 (54.1–61)448/77768 (65.3–70.6)838/123374.9 (70.8–78.8)346/462Surgical urgencyEmergency3.3 (2.6–4.1)68/20

9/132316.4 (13.3–19.8)79/48180.8 (0.4–1.2)16/20560.4 (0–0.8)3/8430.2 (0–0.4)2/13232.3 (1–3.7)11/48190.1 (0–0.2)2/2056NANA0.2 (0–0.5)3/13230.2 (0–0.6)1/481Multimorbidity(≥2 comorbidities)66.2 (63.9–68.3)1216/183757.7 (54.1–61)448/77768 (65.3–70.6)838/123374.9 (70.8–78.8)346/462Surgical urgencyEmergency3.3 (2.6–4.1)68/20 723.6 (2.4–5.1)31/8500.8 (0.4–1.3)11/1337NANAUrgent13.4 (11.9–14.9)278/20728.5 (6.6–10.2)72/85010.7 (9.1–12.3)143/1337NANAExpedited56.2 (54.1–58.3)1164/207273.2 (70.1–76.1)622/85082.9 (80.9–84.7)1108/1337NANAPlanned27.1 (25.1–29.1)562/207214.7 (12.2–17.1)125/8505.6 (4.4–6.8)75/1337NANA Daycase17.7 (16.1–19.4)367/207230.9 (27.9–34.1)263/85047.1 (44.4–49.8)630/1337NANATable 3Postoperative outcomes by subgroups: 85 yr and older, 60–84 yr, daycase surgery, inpatient surgery, elective and non elective surgery.

84.7)1108/1337NANAPlanned27.1 (25.1–29.1)562/207214.7 (12.2–17.1)125/8505.6 (4.4–6.8)75/1337NANA Daycase17.7 (16.1–19.4)367/207230.9 (27.9–34.1)263/85047.1 (44.4–49.8)630/1337NANATable 3Postoperative outcomes by subgroups: 85 yr and older, 60–84 yr, daycase surgery, inpatient surgery, elective and non elective surgery. LOS, delirium within 7 days of surgery, complications (excluding delirium) within 7 days of surgery, mortality at 30 days, 120 days, and 1 yr. Values are percentages with 95% confidence intervals (CI) except for median LOS which is measured in days with interquartile range. Confidence intervals are calculated using the percentile method with 2000 bootstraps. Missing data are omitted from this table and are reported in Supplementary Table S16. LOS, length of stay, ∗ postoperative morbidity excluding delirium.Table 3Characteristic≥85 Yr<85 YrDay caseInpatientElectiveNon-electiveSummaryNumberSummaryNumberSummaryNumberSummaryNumberSummaryNumberSummaryNumberPostoperative LOS (days)5 (1–14)7201 (0–4)61810 (0–0)23643 (1–7)45401 (0–3)48235 (1–13)2079Delirium22.1 (19–25.2)165/7475 (4.5–5.5)316/63090.1 (0–0.2)2/240710.2 (9.3–11)480/47272.7 (2.3–3.2)134/492216.3 (14.7–17.9)348/2138Postoperative morbidity ∗49.1 (45.4–52.9)367/74724.3 (23.3–25.5)1534/63090.2 (0–0.3)4/240740.2 (38.8–41.6)1899/472717.2 (16.1–18.2)845/492249.5 (47.4–51.6)1058/2138Mortality (30 days)3.9 (2.6–5.3)29/7410.8 (0.6–1)49/62880.2 (0–0.4)5/24021.6 (1.2–1.9)73/46330.3 (0.1–0.4)14/49063 (2.3–3.8)64/2124Mortality (120 days)9.9 (7.7–11.9)73/7412.3 (1.9–2.6)142/62880.5 (0.3–0.8)13/24024.4 (3.8–4.9)202/46331.1 (0.8–1.4)54/49067.6 (6.5–8.7)161/2124Mortality (1 yr)18.2 (15.5–21.2)135/7415.6 (5–6.2)353/62882.8 (2.2–3.5)67/24029.1 (8.3–9.9)421/46333.8 (3.3–4.4)187/490614.2 (12.7–15.6)301/2124Table 4Postoperative outcomes by subgroups: orthopaedic, colorectal, urological, and hip fracture surgery. LOS, delirium within 7 days of surgery, complications (excluding delirium) within 7 days of surgery, mortality at 30 days, 120 days, and 1 yr. Values are percentages with 95% confidence intervals (CI) except for median LOS which is measured in days with interquartile range. Confidence intervals are calculated using the percentile method with 2000 bootstraps. Missing data are omitted from this table and are reported in Supplementary Table S16. LOS, length of stay.

es are percentages with 95% confidence intervals (CI) except for median LOS which is measured in days with interquartile range. Confidence intervals are calculated using the percentile method with 2000 bootstraps. Missing data are omitted from this table and are reported in Supplementary Table S16. LOS, length of stay. ∗, postoperative morbidity excluding delirium.Table 4CharacteristicOrthopaedicsColorectalUrologyHip fractureSummaryNumberSummaryNumberSummaryNumberSummaryNumberPostoperative LOS (days)3 (1–9)20272 (0–6)8361 (0–1)131613 (7–20)465Delirium13.1 (11.7–14.5)271/20725.3 (3.8–6.8)45/8501.9 (1.2–2.6)25/133736.8 (32.4–40.9)178/484Postoperative morbidity∗42.3 (40.1–44.3)876/207232.4 (29.3–35.4)275/8509.5 (7.9–11.1)127/133785.3 (82–88.4)413/484Mortality (30 days)1.8 (1.3–2.4)37/20590.9 (0.4–1.7)8/8470.7 (0.3–1.2)9/13345 (3.1–7.1)24/480Mortality (120 days)4.5 (3.6–5.4)93/20592.7 (1.7–3.9)23/8471.9 (1.2–2.7)25/133412.5 (9.6–15.4)60/480Mortality (1 yr)8.6 (7.3–9.8)177/20596.3 (4.7–7.9)53/8475.8 (4.6–7.2)77/133421.9 (18.1–25.8)105/480

27/133785.3 (82–88.4)413/484Mortality (30 days)1.8 (1.3–2.4)37/20590.9 (0.4–1.7)8/8470.7 (0.3–1.2)9/13345 (3.1–7.1)24/480Mortality (120 days)4.5 (3.6–5.4)93/20592.7 (1.7–3.9)23/8471.9 (1.2–2.7)25/133412.5 (9.6–15.4)60/480Mortality (1 yr)8.6 (7.3–9.8)177/20596.3 (4.7–7.9)53/8475.8 (4.6–7.2)77/133421.9 (18.1–25.8)105/480 Participant characteristics of the subgroups: 85 yr and older, 60–84 yr, daycase surgery, inpatient surgery, elective, and non elective surgery. The participant characteristics of the SNAP-3 subgroups. Percentages have been rounded so may not total 100% exactly. Missing data are omitted from this table and reported in Supplementary Table S9. Surgical urgency is defined using NCEPOD categorisations.15 The values here are proportions (%) with 95% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps. Surgical urgency is described using the NCEPOD criteria.15 ASA, American Society of Anesthesiologists; NA, not applicable; NCEPOD, National Confidential Enquiry into Patient Outcome and Death.

% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps. Surgical urgency is described using the NCEPOD criteria.15 ASA, American Society of Anesthesiologists; NA, not applicable; NCEPOD, National Confidential Enquiry into Patient Outcome and Death. Participant characteristics of the subgroups: orthopaedic, colorectal, urological, and hip fracture surgery. The participant characteristics of the SNAP-3 subgroups. Percentages have been rounded so may not total 100% exactly. Missing data are omitted from this table and reported in Supplementary Table S9. Surgical urgency is defined using NCEPOD categorisations.15 The values here are proportions (%) with 95% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps. Surgical urgency is described using the NCEPOD criteria.15 ASA, American Society of Anesthesiologists; NA, not applicable; NCEPOD, National Confidential Enquiry into Patient Outcome and Death.

% confidence intervals (95% CI) apart from mean age. The 95% CI are calculated using the percentile method and 2000 bootstraps. Surgical urgency is described using the NCEPOD criteria.15 ASA, American Society of Anesthesiologists; NA, not applicable; NCEPOD, National Confidential Enquiry into Patient Outcome and Death. Postoperative outcomes by subgroups: 85 yr and older, 60–84 yr, daycase surgery, inpatient surgery, elective and non elective surgery. LOS, delirium within 7 days of surgery, complications (excluding delirium) within 7 days of surgery, mortality at 30 days, 120 days, and 1 yr. Values are percentages with 95% confidence intervals (CI) except for median LOS which is measured in days with interquartile range. Confidence intervals are calculated using the percentile method with 2000 bootstraps. Missing data are omitted from this table and are reported in Supplementary Table S16. LOS, length of stay, ∗ postoperative morbidity excluding delirium. Postoperative outcomes by subgroups: orthopaedic, colorectal, urological, and hip fracture surgery. LOS, delirium within 7 days of surgery, complications (excluding delirium) within 7 days of surgery, mortality at 30 days, 120 days, and 1 yr. Values are percentages with 95% confidence intervals (CI) except for median LOS which is measured in days with interquartile range. Confidence intervals are calculated using the percentile method with 2000 bootstraps. Missing data are omitted from this table and are reported in Supplementary Table S16. LOS, length of stay. ∗, postoperative morbidity excluding delirium.

s (CI) except for median LOS which is measured in days with interquartile range. Confidence intervals are calculated using the percentile method with 2000 bootstraps. Missing data are omitted from this table and are reported in Supplementary Table S16. LOS, length of stay. ∗, postoperative morbidity excluding delirium. The SNAP-3 cohort contained 747 participants aged ≥85 yr, making up 8.6% of the overall cohort, with 3.3% (231/7134) aged ≥90 yr. About half (365/740) of those aged ≥85 yr were living with frailty, in comparison with 16.0% (1003/6252) of younger participants (Fig. 1). Of those aged ≥85 yr, 80.6% (561/696) lived with multimorbidity (Fig. 1), compared with 60.9% (3415/5604) of younger participants. Around half (343/688) of those aged ≥85 yr required help with instrumental ADLs, 7.6% (56/688) needed assistance with basic ADLs, and 85.7% (640/747) lived in their own home.Fig 1Prevalence of frailty and multimorbidity across subgroups of older surgical patients. The prevalence (with 95% confidence intervals) of frailty and multimorbidity in older surgical patients within SNAP-3. Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.Fig 1 Prevalence of frailty and multimorbidity across subgroups of older surgical patients. The prevalence (with 95% confidence intervals) of frailty and multimorbidity in older surgical patients within SNAP-3. Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.

The SNAP-3 cohort contained 747 participants aged ≥85 yr, making up 8.6% of the overall cohort, with 3.3% (231/7134) aged ≥90 yr. About half (365/740) of those aged ≥85 yr were living with frailty, in comparison with 16.0% (1003/6252) of younger participants (Fig. 1). Of those aged ≥85 yr, 80.6% (561/696) lived with multimorbidity (Fig. 1), compared with 60.9% (3415/5604) of younger participants. Around half (343/688) of those aged ≥85 yr required help with instrumental ADLs, 7.6% (56/688) needed assistance with basic ADLs, and 85.7% (640/747) lived in their own home.Fig 1Prevalence of frailty and multimorbidity across subgroups of older surgical patients. The prevalence (with 95% confidence intervals) of frailty and multimorbidity in older surgical patients within SNAP-3. Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.Fig 1 Prevalence of frailty and multimorbidity across subgroups of older surgical patients. The prevalence (with 95% confidence intervals) of frailty and multimorbidity in older surgical patients within SNAP-3. Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval. Almost half (358/742) of those aged ≥85 yr were undergoing orthopaedic procedures, with urology and colorectal being the next most common surgical specialties. Procedures were elective in 45% (366/747) of all cases carried out on those aged ≥85 yr.

Prevalence of frailty and multimorbidity across subgroups of older surgical patients. The prevalence (with 95% confidence intervals) of frailty and multimorbidity in older surgical patients within SNAP-3. Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval. Almost half (358/742) of those aged ≥85 yr were undergoing orthopaedic procedures, with urology and colorectal being the next most common surgical specialties. Procedures were elective in 45% (366/747) of all cases carried out on those aged ≥85 yr. Those aged 85 yr or older had a median postoperative hospital stay which was five times longer than younger participants (5 [IQR 1–14] days vs 1 [IQR 0–4] days). The observed incidence of postoperative delirium in those aged ≥85 yr (165/747; 22.1%) was four times higher than in those aged 60–84 yr (316/6309; 5%). Similarly, the incidence of non-delirium postoperative complications in those aged ≥85 yr (367/747; 49.1%) was double that of those aged 60–84 yr (1534/6309; 24.3%) (Fig. 2). Mortality at 30 days was five times higher in those aged ≥85 yr (29/741; 4.3%) compared with those aged 60–84 yr (49/6288; 0.8%) (Fig. 3).Fig 2Incidence of postoperative outcomes across subgroups of older surgical patients. The incidence of postoperative outcomes in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. Delirium includes any postoperative delirium within 7 days of surgery and morbidity includes any postoperative complications excluding delirium within 7 days of surgery. CI, confidence interval.Fig 2Fig 3Mortality at 30 days, 120 days, and 1 yr after surgery across subgroups of older surgical patients. The incidence of postoperative mortality in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.Fig 3

at 30 days, 120 days, and 1 yr after surgery across subgroups of older surgical patients. The incidence of postoperative mortality in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.Fig 3 Incidence of postoperative outcomes across subgroups of older surgical patients. The incidence of postoperative outcomes in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. Delirium includes any postoperative delirium within 7 days of surgery and morbidity includes any postoperative complications excluding delirium within 7 days of surgery. CI, confidence interval. Mortality at 30 days, 120 days, and 1 yr after surgery across subgroups of older surgical patients. The incidence of postoperative mortality in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval. One-third (2364/7134) of all operations were daycase procedures, of which 84.5% (2032/2406) were elective.

Mortality at 30 days, 120 days, and 1 yr after surgery across subgroups of older surgical patients. The incidence of postoperative mortality in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval. One-third (2364/7134) of all operations were daycase procedures, of which 84.5% (2032/2406) were elective. The use of daycase surgery declined with advancing age; 16.0% (385/2407) of daycase procedures involved individuals aged ≥80 yr, compared with 26.9% (1247/4650) of inpatient procedures. Of those aged ≥85 yr, 17.5% (138/747) of procedures were daycase, compared with 35.9% (2268/6309) of younger participants. The day surgery cohorts were less often living with multimorbidity and had lower ASA physical status scores than their inpatient counterparts (Table 1). The prevalence of frailty in the inpatient cohort (1116/4617; 24.2%) was twice that of the daycase cohort (253/2127; 12.4%) (Fig. 1). The most common non elective day surgery cases included open reduction internal fixation (ORIF) of a long bone and transurethral resection of a bladder tumour (TURBT). Mortality rates at 30 days were approximately eight times higher in those requiring inpatient surgery (73/4633; 1.6%) than those having day surgery (5/2402; 0.2%) and three times higher at 1 yr (421/4633 [9.1%] vs 67/2402 [2.8%]). Participants undergoing elective procedures made up 69% (4922/7134) of the SNAP-3 cohort.

The most common non elective day surgery cases included open reduction internal fixation (ORIF) of a long bone and transurethral resection of a bladder tumour (TURBT). Mortality rates at 30 days were approximately eight times higher in those requiring inpatient surgery (73/4633; 1.6%) than those having day surgery (5/2402; 0.2%) and three times higher at 1 yr (421/4633 [9.1%] vs 67/2402 [2.8%]). Participants undergoing elective procedures made up 69% (4922/7134) of the SNAP-3 cohort. Elective patients were younger than the non elective subgroup, with 4004/4922 (81.3%) of elective participants <80 yr old, compared with 1419/2138 (66.4%) of non elective participants. The prevalence of frailty in the elective subgroup (701/4870; 14.4%) was half that of the non elective subgroup (668/2126; 31.4%). Multimorbidity prevalence was similar in those undergoing elective (2650/4339; 61.1%) and non elective surgery (1328/1964; 67.6%) (Fig. 1). The prevalence of frailty increased with increasing urgency of surgery from elective, expedited to urgent surgery. Participants undergoing elective surgery were more often independent with ADLs (3739/4839; 77.3%) than those undergoing non elective surgery (1252/2007; 62.4%). They were also less likely to be admitted from residential care (0.4%; 22/4919 vs 4.2%; 89/2137) or lack capacity to consent for this research study (0.8%; 39/4922 vs 9.4%; 201/2138) than those undergoing non elective surgery.

t with ADLs (3739/4839; 77.3%) than those undergoing non elective surgery (1252/2007; 62.4%). They were also less likely to be admitted from residential care (0.4%; 22/4919 vs 4.2%; 89/2137) or lack capacity to consent for this research study (0.8%; 39/4922 vs 9.4%; 201/2138) than those undergoing non elective surgery. Orthopaedic surgery was the most common surgical specialty for both elective surgery (1164/4852; 24%) and non elective surgery (908/2112; 43%), followed by urology and colorectal (elective) and upper gastrointestinal and colorectal procedures (non elective). The median postoperative length of stay for those undergoing non elective surgery (5 [IQR 1–13] days) was five times that of an elective patient (1 [IQR 0–3] days). The incidence of postoperative delirium was six times higher in the non elective group (16.3%; 348/2138) than the elective group (2.7%; 134/4922). Postoperative morbidity within 7 days (excluding delirium) was almost triple in the non elective participants (149.5%; 058/2138) compared with the elective participants (17.2%; 845/4922) (Fig. 2). Mortality after non elective surgery was greater at all time points compared with elective surgery (Fig. 3). Orthopaedic surgery was the predominant surgical specialty in the SNAP-3 study, comprising nearly one in three (29.0%; 2072/7134) procedures in older surgical patients. Urology (19.2%; 1337/7134) and colorectal surgery (12.2%; 850/7134) were the next most prevalent surgical specialities.

The median postoperative length of stay for those undergoing non elective surgery (5 [IQR 1–13] days) was five times that of an elective patient (1 [IQR 0–3] days). The incidence of postoperative delirium was six times higher in the non elective group (16.3%; 348/2138) than the elective group (2.7%; 134/4922). Postoperative morbidity within 7 days (excluding delirium) was almost triple in the non elective participants (149.5%; 058/2138) compared with the elective participants (17.2%; 845/4922) (Fig. 2). Mortality after non elective surgery was greater at all time points compared with elective surgery (Fig. 3). Orthopaedic surgery was the predominant surgical specialty in the SNAP-3 study, comprising nearly one in three (29.0%; 2072/7134) procedures in older surgical patients. Urology (19.2%; 1337/7134) and colorectal surgery (12.2%; 850/7134) were the next most prevalent surgical specialities. Males comprised 77.5% (1036/1337) of those undergoing urological procedures and 63.1% (314/850) of those undergoing colorectal procedures. In contrast, females were more prevalent in the orthopaedic cohort, making up 58.7% (1217/2072) of participants. The mean age and BMI across these three specialties was similar; more than a quarter had class 1 obesity or greater. ASA physical status was also comparable across these specialties, with the exception of a greater proportion of orthopaedic participants classified as ASA physical status 4 (5.2%; 107/2059) compared with those undergoing non-orthopaedic surgery (2.2%; 49/2169).

milar; more than a quarter had class 1 obesity or greater. ASA physical status was also comparable across these specialties, with the exception of a greater proportion of orthopaedic participants classified as ASA physical status 4 (5.2%; 107/2059) compared with those undergoing non-orthopaedic surgery (2.2%; 49/2169). Frailty was more prevalent in the orthopaedic cohort, with nearly one-third (30.6%; 630/2056) of participants living with frailty, compared with nearly one-sixth (14.5%; 315/2166) of those undergoing non-orthopaedic surgery (Fig. 1). Participants undergoing orthopaedic surgery were also less likely to be independent in ADLs. Amongst this group, 32.8% (649/1978) required assistance with instrumental ADLs, and 6.0% (119/1978) needed help with basic ADLs. Multimorbidity was more prevalent in participants undergoing orthopaedic (66.2%; 1216/1837) and urological (68.0%; 838/1233) procedures than colorectal surgery (57.7%; 448/777) (Fig. 1). The differences in frailty and functional status between orthopaedics and other specialties were most marked in the non elective participants, but still evident in elective participants (Supplementary Table S9). Orthopaedic surgery had the highest proportion of non elective participants 43.8%; (908/2072) and the lowest proportion of day surgery procedures (17.7%; 367/2072). Urology had the highest prevalence of day surgery procedures (47.1%; 630/1337).

Frailty was more prevalent in the orthopaedic cohort, with nearly one-third (30.6%; 630/2056) of participants living with frailty, compared with nearly one-sixth (14.5%; 315/2166) of those undergoing non-orthopaedic surgery (Fig. 1). Participants undergoing orthopaedic surgery were also less likely to be independent in ADLs. Amongst this group, 32.8% (649/1978) required assistance with instrumental ADLs, and 6.0% (119/1978) needed help with basic ADLs. Multimorbidity was more prevalent in participants undergoing orthopaedic (66.2%; 1216/1837) and urological (68.0%; 838/1233) procedures than colorectal surgery (57.7%; 448/777) (Fig. 1). The differences in frailty and functional status between orthopaedics and other specialties were most marked in the non elective participants, but still evident in elective participants (Supplementary Table S9). Orthopaedic surgery had the highest proportion of non elective participants 43.8%; (908/2072) and the lowest proportion of day surgery procedures (17.7%; 367/2072). Urology had the highest prevalence of day surgery procedures (47.1%; 630/1337). Patients undergoing elective orthopaedic, urology, or colorectal surgery had similar median postoperative lengths of stay: orthopaedics, 2 (IQR 1–4) days; urology, 1 (IQR 0–1) day; and colorectal, 1 (IQR 1–9) day. In contrast, among non elective patients, there was greater variation in postoperative length of stay. Non-elective orthopaedic patients had a median stay of 9 (IQR 3–18) days, colorectal patients 5 (IQR 1–9) days, and non elective urology 1 (IQR 0–2) day.

ays; urology, 1 (IQR 0–1) day; and colorectal, 1 (IQR 1–9) day. In contrast, among non elective patients, there was greater variation in postoperative length of stay. Non-elective orthopaedic patients had a median stay of 9 (IQR 3–18) days, colorectal patients 5 (IQR 1–9) days, and non elective urology 1 (IQR 0–2) day. Delirium was most prevalent in the non elective orthopaedic cohort (25.2%; 229/908), less frequently affecting non elective colorectal (11.0%; 25/228) and non elective urology participants (5.2%; 12/229). The incidences of delirium in the elective cohorts of orthopaedic and colorectal participants were similar at 3.6% (42/1164) and 3.2% (20/622), respectively, with elective urology participants least often experiencing delirium with an incidence of 1.2% (13/1108). Other postoperative complications were also more common amongst non elective orthopaedic patients (65.7%; 597/908) than colorectal (48.2%; 110/228) and urology patients (17%; 39/229) (Fig. 2). Mortality rates at 30 days were similar between surgical specialities, varying depending on whether they were undergoing elective or non elective surgery. Those undergoing elective colorectal and urology surgery had a higher incidence of 120-day mortality: colorectal 1.9% (12/620), urology 1.0% (11/1105); and 1-yr mortality: colorectal 4.5% (28/620), urology 4.3% (48/1105) than elective orthopaedics (120-day mortality 0.8% [9/1162] and 1-yr mortality 2.1% [25/1162]) (Fig. 3). Participants undergoing surgery for hip fracture comprised 6.8% (484/7134) of the SNAP-3 cohort.

Mortality rates at 30 days were similar between surgical specialities, varying depending on whether they were undergoing elective or non elective surgery. Those undergoing elective colorectal and urology surgery had a higher incidence of 120-day mortality: colorectal 1.9% (12/620), urology 1.0% (11/1105); and 1-yr mortality: colorectal 4.5% (28/620), urology 4.3% (48/1105) than elective orthopaedics (120-day mortality 0.8% [9/1162] and 1-yr mortality 2.1% [25/1162]) (Fig. 3). Participants undergoing surgery for hip fracture comprised 6.8% (484/7134) of the SNAP-3 cohort. Amongst hip fracture participants, 69.8% (338/484) were female, with a mean age of 81.6 (sd 9.0) yr. In the non-hip fracture group, 47.6% were female, and the mean age was younger at 72.2 (sd 7.8) yr. Participants with hip fracture had lower BMI compared with the rest of the SNAP-3 cohort. In those with hip fracture, 53.3% (250/469) were of normal weight, and 9.2% (43/469) were underweight. In those without hip fracture, 28.2% (1846/6531) were of normal weight, and 1.5% were underweight (95/6531). The ASA physical status was higher in patients with hip fracture compared with those without: amongst those with hip fracture 70.2% (339/483) were ASA physical status 3 or 4, compared with 37% (2412/6516) of those without hip fracture.

cture, 28.2% (1846/6531) were of normal weight, and 1.5% were underweight (95/6531). The ASA physical status was higher in patients with hip fracture compared with those without: amongst those with hip fracture 70.2% (339/483) were ASA physical status 3 or 4, compared with 37% (2412/6516) of those without hip fracture. Of those undergoing surgery for hip fracture, 25.4% (123/484) did not have capacity to consent to this study, with 19.4% (94/484) having dementia. Two-thirds needed assistance with ADLs (help with instrumental ADLs 53.9% [229/425] and basic ADLs 10.8% [46/425]). Only 35.3% (150/425) were independent, despite 80.8% (391/484) being admitted from their own home. Polypharmacy was common with a prevalence of 58.8% (280/476). The prevalence of frailty was notably higher amongst patients with hip fracture; 58.2% (280/481) were living with frailty, compared with 16.7% (1089/6516) of those without hip fracture. Multimorbidity was also more common in patients with hip fracture, affecting 74.9% (346/462), compared with 62.2% (3632/5841) of those without hip fracture (Fig. 1). Median postoperative length of stay was 13 (IQR 7–20) days for those with hip fracture. The incidence of postoperative delirium was 36.8% (178/484), and of other complications was 85.3% (413/484) (Fig. 2). Mortality rates were higher amongst patients with hip fracture compared with those without, with a 5.0% (24/480) 30-day mortality, 12.5% (60/480) 120-day mortality, and 21.7% (105/480) 1-yr mortality rate (Fig. 3).

at 30 days, 120 days, and 1 yr after surgery across subgroups of older surgical patients. The incidence of postoperative mortality in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.Fig 3 Incidence of postoperative outcomes across subgroups of older surgical patients. The incidence of postoperative outcomes in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. Delirium includes any postoperative delirium within 7 days of surgery and morbidity includes any postoperative complications excluding delirium within 7 days of surgery. CI, confidence interval. Mortality at 30 days, 120 days, and 1 yr after surgery across subgroups of older surgical patients. The incidence of postoperative mortality in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.

The SNAP-3 cohort contained 747 participants aged ≥85 yr, making up 8.6% of the overall cohort, with 3.3% (231/7134) aged ≥90 yr. About half (365/740) of those aged ≥85 yr were living with frailty, in comparison with 16.0% (1003/6252) of younger participants (Fig. 1). Of those aged ≥85 yr, 80.6% (561/696) lived with multimorbidity (Fig. 1), compared with 60.9% (3415/5604) of younger participants. Around half (343/688) of those aged ≥85 yr required help with instrumental ADLs, 7.6% (56/688) needed assistance with basic ADLs, and 85.7% (640/747) lived in their own home.Fig 1Prevalence of frailty and multimorbidity across subgroups of older surgical patients. The prevalence (with 95% confidence intervals) of frailty and multimorbidity in older surgical patients within SNAP-3. Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.Fig 1 Prevalence of frailty and multimorbidity across subgroups of older surgical patients. The prevalence (with 95% confidence intervals) of frailty and multimorbidity in older surgical patients within SNAP-3. Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.

Almost half (358/742) of those aged ≥85 yr were undergoing orthopaedic procedures, with urology and colorectal being the next most common surgical specialties. Procedures were elective in 45% (366/747) of all cases carried out on those aged ≥85 yr.

Those aged 85 yr or older had a median postoperative hospital stay which was five times longer than younger participants (5 [IQR 1–14] days vs 1 [IQR 0–4] days). The observed incidence of postoperative delirium in those aged ≥85 yr (165/747; 22.1%) was four times higher than in those aged 60–84 yr (316/6309; 5%). Similarly, the incidence of non-delirium postoperative complications in those aged ≥85 yr (367/747; 49.1%) was double that of those aged 60–84 yr (1534/6309; 24.3%) (Fig. 2). Mortality at 30 days was five times higher in those aged ≥85 yr (29/741; 4.3%) compared with those aged 60–84 yr (49/6288; 0.8%) (Fig. 3).Fig 2Incidence of postoperative outcomes across subgroups of older surgical patients. The incidence of postoperative outcomes in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. Delirium includes any postoperative delirium within 7 days of surgery and morbidity includes any postoperative complications excluding delirium within 7 days of surgery. CI, confidence interval.Fig 2Fig 3Mortality at 30 days, 120 days, and 1 yr after surgery across subgroups of older surgical patients. The incidence of postoperative mortality in older surgical patients within SNAP-3 (with 95% confidence intervals). Confidence intervals are calculated using the percentile method with 2000 bootstraps. CI, confidence interval.Fig 3

One-third (2364/7134) of all operations were daycase procedures, of which 84.5% (2032/2406) were elective. The use of daycase surgery declined with advancing age; 16.0% (385/2407) of daycase procedures involved individuals aged ≥80 yr, compared with 26.9% (1247/4650) of inpatient procedures. Of those aged ≥85 yr, 17.5% (138/747) of procedures were daycase, compared with 35.9% (2268/6309) of younger participants. The day surgery cohorts were less often living with multimorbidity and had lower ASA physical status scores than their inpatient counterparts (Table 1). The prevalence of frailty in the inpatient cohort (1116/4617; 24.2%) was twice that of the daycase cohort (253/2127; 12.4%) (Fig. 1). The most common non elective day surgery cases included open reduction internal fixation (ORIF) of a long bone and transurethral resection of a bladder tumour (TURBT). Mortality rates at 30 days were approximately eight times higher in those requiring inpatient surgery (73/4633; 1.6%) than those having day surgery (5/2402; 0.2%) and three times higher at 1 yr (421/4633 [9.1%] vs 67/2402 [2.8%]).

The use of daycase surgery declined with advancing age; 16.0% (385/2407) of daycase procedures involved individuals aged ≥80 yr, compared with 26.9% (1247/4650) of inpatient procedures. Of those aged ≥85 yr, 17.5% (138/747) of procedures were daycase, compared with 35.9% (2268/6309) of younger participants. The day surgery cohorts were less often living with multimorbidity and had lower ASA physical status scores than their inpatient counterparts (Table 1). The prevalence of frailty in the inpatient cohort (1116/4617; 24.2%) was twice that of the daycase cohort (253/2127; 12.4%) (Fig. 1).

Mortality rates at 30 days were approximately eight times higher in those requiring inpatient surgery (73/4633; 1.6%) than those having day surgery (5/2402; 0.2%) and three times higher at 1 yr (421/4633 [9.1%] vs 67/2402 [2.8%]).

Participants undergoing elective procedures made up 69% (4922/7134) of the SNAP-3 cohort. Elective patients were younger than the non elective subgroup, with 4004/4922 (81.3%) of elective participants <80 yr old, compared with 1419/2138 (66.4%) of non elective participants. The prevalence of frailty in the elective subgroup (701/4870; 14.4%) was half that of the non elective subgroup (668/2126; 31.4%). Multimorbidity prevalence was similar in those undergoing elective (2650/4339; 61.1%) and non elective surgery (1328/1964; 67.6%) (Fig. 1). The prevalence of frailty increased with increasing urgency of surgery from elective, expedited to urgent surgery. Participants undergoing elective surgery were more often independent with ADLs (3739/4839; 77.3%) than those undergoing non elective surgery (1252/2007; 62.4%). They were also less likely to be admitted from residential care (0.4%; 22/4919 vs 4.2%; 89/2137) or lack capacity to consent for this research study (0.8%; 39/4922 vs 9.4%; 201/2138) than those undergoing non elective surgery. Orthopaedic surgery was the most common surgical specialty for both elective surgery (1164/4852; 24%) and non elective surgery (908/2112; 43%), followed by urology and colorectal (elective) and upper gastrointestinal and colorectal procedures (non elective).

The prevalence of frailty in the elective subgroup (701/4870; 14.4%) was half that of the non elective subgroup (668/2126; 31.4%). Multimorbidity prevalence was similar in those undergoing elective (2650/4339; 61.1%) and non elective surgery (1328/1964; 67.6%) (Fig. 1). The prevalence of frailty increased with increasing urgency of surgery from elective, expedited to urgent surgery. Participants undergoing elective surgery were more often independent with ADLs (3739/4839; 77.3%) than those undergoing non elective surgery (1252/2007; 62.4%). They were also less likely to be admitted from residential care (0.4%; 22/4919 vs 4.2%; 89/2137) or lack capacity to consent for this research study (0.8%; 39/4922 vs 9.4%; 201/2138) than those undergoing non elective surgery. Orthopaedic surgery was the most common surgical specialty for both elective surgery (1164/4852; 24%) and non elective surgery (908/2112; 43%), followed by urology and colorectal (elective) and upper gastrointestinal and colorectal procedures (non elective). The median postoperative length of stay for those undergoing non elective surgery (5 [IQR 1–13] days) was five times that of an elective patient (1 [IQR 0–3] days). The incidence of postoperative delirium was six times higher in the non elective group (16.3%; 348/2138) than the elective group (2.7%; 134/4922). Postoperative morbidity within 7 days (excluding delirium) was almost triple in the non elective participants (149.5%; 058/2138) compared with the elective participants (17.2%; 845/4922) (Fig. 2). Mortality after non elective surgery was greater at all time points compared with elective surgery (Fig. 3).

Elective patients were younger than the non elective subgroup, with 4004/4922 (81.3%) of elective participants <80 yr old, compared with 1419/2138 (66.4%) of non elective participants. The prevalence of frailty in the elective subgroup (701/4870; 14.4%) was half that of the non elective subgroup (668/2126; 31.4%). Multimorbidity prevalence was similar in those undergoing elective (2650/4339; 61.1%) and non elective surgery (1328/1964; 67.6%) (Fig. 1). The prevalence of frailty increased with increasing urgency of surgery from elective, expedited to urgent surgery. Participants undergoing elective surgery were more often independent with ADLs (3739/4839; 77.3%) than those undergoing non elective surgery (1252/2007; 62.4%). They were also less likely to be admitted from residential care (0.4%; 22/4919 vs 4.2%; 89/2137) or lack capacity to consent for this research study (0.8%; 39/4922 vs 9.4%; 201/2138) than those undergoing non elective surgery.

Orthopaedic surgery was the most common surgical specialty for both elective surgery (1164/4852; 24%) and non elective surgery (908/2112; 43%), followed by urology and colorectal (elective) and upper gastrointestinal and colorectal procedures (non elective).

The median postoperative length of stay for those undergoing non elective surgery (5 [IQR 1–13] days) was five times that of an elective patient (1 [IQR 0–3] days). The incidence of postoperative delirium was six times higher in the non elective group (16.3%; 348/2138) than the elective group (2.7%; 134/4922). Postoperative morbidity within 7 days (excluding delirium) was almost triple in the non elective participants (149.5%; 058/2138) compared with the elective participants (17.2%; 845/4922) (Fig. 2). Mortality after non elective surgery was greater at all time points compared with elective surgery (Fig. 3).

Orthopaedic surgery was the predominant surgical specialty in the SNAP-3 study, comprising nearly one in three (29.0%; 2072/7134) procedures in older surgical patients. Urology (19.2%; 1337/7134) and colorectal surgery (12.2%; 850/7134) were the next most prevalent surgical specialities. Males comprised 77.5% (1036/1337) of those undergoing urological procedures and 63.1% (314/850) of those undergoing colorectal procedures. In contrast, females were more prevalent in the orthopaedic cohort, making up 58.7% (1217/2072) of participants. The mean age and BMI across these three specialties was similar; more than a quarter had class 1 obesity or greater. ASA physical status was also comparable across these specialties, with the exception of a greater proportion of orthopaedic participants classified as ASA physical status 4 (5.2%; 107/2059) compared with those undergoing non-orthopaedic surgery (2.2%; 49/2169).

ays; urology, 1 (IQR 0–1) day; and colorectal, 1 (IQR 1–9) day. In contrast, among non elective patients, there was greater variation in postoperative length of stay. Non-elective orthopaedic patients had a median stay of 9 (IQR 3–18) days, colorectal patients 5 (IQR 1–9) days, and non elective urology 1 (IQR 0–2) day. Delirium was most prevalent in the non elective orthopaedic cohort (25.2%; 229/908), less frequently affecting non elective colorectal (11.0%; 25/228) and non elective urology participants (5.2%; 12/229). The incidences of delirium in the elective cohorts of orthopaedic and colorectal participants were similar at 3.6% (42/1164) and 3.2% (20/622), respectively, with elective urology participants least often experiencing delirium with an incidence of 1.2% (13/1108). Other postoperative complications were also more common amongst non elective orthopaedic patients (65.7%; 597/908) than colorectal (48.2%; 110/228) and urology patients (17%; 39/229) (Fig. 2). Mortality rates at 30 days were similar between surgical specialities, varying depending on whether they were undergoing elective or non elective surgery. Those undergoing elective colorectal and urology surgery had a higher incidence of 120-day mortality: colorectal 1.9% (12/620), urology 1.0% (11/1105); and 1-yr mortality: colorectal 4.5% (28/620), urology 4.3% (48/1105) than elective orthopaedics (120-day mortality 0.8% [9/1162] and 1-yr mortality 2.1% [25/1162]) (Fig. 3).

Males comprised 77.5% (1036/1337) of those undergoing urological procedures and 63.1% (314/850) of those undergoing colorectal procedures. In contrast, females were more prevalent in the orthopaedic cohort, making up 58.7% (1217/2072) of participants. The mean age and BMI across these three specialties was similar; more than a quarter had class 1 obesity or greater. ASA physical status was also comparable across these specialties, with the exception of a greater proportion of orthopaedic participants classified as ASA physical status 4 (5.2%; 107/2059) compared with those undergoing non-orthopaedic surgery (2.2%; 49/2169). Frailty was more prevalent in the orthopaedic cohort, with nearly one-third (30.6%; 630/2056) of participants living with frailty, compared with nearly one-sixth (14.5%; 315/2166) of those undergoing non-orthopaedic surgery (Fig. 1). Participants undergoing orthopaedic surgery were also less likely to be independent in ADLs. Amongst this group, 32.8% (649/1978) required assistance with instrumental ADLs, and 6.0% (119/1978) needed help with basic ADLs. Multimorbidity was more prevalent in participants undergoing orthopaedic (66.2%; 1216/1837) and urological (68.0%; 838/1233) procedures than colorectal surgery (57.7%; 448/777) (Fig. 1). The differences in frailty and functional status between orthopaedics and other specialties were most marked in the non elective participants, but still evident in elective participants (Supplementary Table S9).

Orthopaedic surgery had the highest proportion of non elective participants 43.8%; (908/2072) and the lowest proportion of day surgery procedures (17.7%; 367/2072). Urology had the highest prevalence of day surgery procedures (47.1%; 630/1337).

Patients undergoing elective orthopaedic, urology, or colorectal surgery had similar median postoperative lengths of stay: orthopaedics, 2 (IQR 1–4) days; urology, 1 (IQR 0–1) day; and colorectal, 1 (IQR 1–9) day. In contrast, among non elective patients, there was greater variation in postoperative length of stay. Non-elective orthopaedic patients had a median stay of 9 (IQR 3–18) days, colorectal patients 5 (IQR 1–9) days, and non elective urology 1 (IQR 0–2) day. Delirium was most prevalent in the non elective orthopaedic cohort (25.2%; 229/908), less frequently affecting non elective colorectal (11.0%; 25/228) and non elective urology participants (5.2%; 12/229). The incidences of delirium in the elective cohorts of orthopaedic and colorectal participants were similar at 3.6% (42/1164) and 3.2% (20/622), respectively, with elective urology participants least often experiencing delirium with an incidence of 1.2% (13/1108). Other postoperative complications were also more common amongst non elective orthopaedic patients (65.7%; 597/908) than colorectal (48.2%; 110/228) and urology patients (17%; 39/229) (Fig. 2).

0/622), respectively, with elective urology participants least often experiencing delirium with an incidence of 1.2% (13/1108). Other postoperative complications were also more common amongst non elective orthopaedic patients (65.7%; 597/908) than colorectal (48.2%; 110/228) and urology patients (17%; 39/229) (Fig. 2). Mortality rates at 30 days were similar between surgical specialities, varying depending on whether they were undergoing elective or non elective surgery. Those undergoing elective colorectal and urology surgery had a higher incidence of 120-day mortality: colorectal 1.9% (12/620), urology 1.0% (11/1105); and 1-yr mortality: colorectal 4.5% (28/620), urology 4.3% (48/1105) than elective orthopaedics (120-day mortality 0.8% [9/1162] and 1-yr mortality 2.1% [25/1162]) (Fig. 3).

Participants undergoing surgery for hip fracture comprised 6.8% (484/7134) of the SNAP-3 cohort. Amongst hip fracture participants, 69.8% (338/484) were female, with a mean age of 81.6 (sd 9.0) yr. In the non-hip fracture group, 47.6% were female, and the mean age was younger at 72.2 (sd 7.8) yr. Participants with hip fracture had lower BMI compared with the rest of the SNAP-3 cohort. In those with hip fracture, 53.3% (250/469) were of normal weight, and 9.2% (43/469) were underweight. In those without hip fracture, 28.2% (1846/6531) were of normal weight, and 1.5% were underweight (95/6531). The ASA physical status was higher in patients with hip fracture compared with those without: amongst those with hip fracture 70.2% (339/483) were ASA physical status 3 or 4, compared with 37% (2412/6516) of those without hip fracture. Of those undergoing surgery for hip fracture, 25.4% (123/484) did not have capacity to consent to this study, with 19.4% (94/484) having dementia. Two-thirds needed assistance with ADLs (help with instrumental ADLs 53.9% [229/425] and basic ADLs 10.8% [46/425]). Only 35.3% (150/425) were independent, despite 80.8% (391/484) being admitted from their own home. Polypharmacy was common with a prevalence of 58.8% (280/476). The prevalence of frailty was notably higher amongst patients with hip fracture; 58.2% (280/481) were living with frailty, compared with 16.7% (1089/6516) of those without hip fracture. Multimorbidity was also more common in patients with hip fracture, affecting 74.9% (346/462), compared with 62.2% (3632/5841) of those without hip fracture (Fig. 1).

otably higher amongst patients with hip fracture; 58.2% (280/481) were living with frailty, compared with 16.7% (1089/6516) of those without hip fracture. Multimorbidity was also more common in patients with hip fracture, affecting 74.9% (346/462), compared with 62.2% (3632/5841) of those without hip fracture (Fig. 1). Median postoperative length of stay was 13 (IQR 7–20) days for those with hip fracture. The incidence of postoperative delirium was 36.8% (178/484), and of other complications was 85.3% (413/484) (Fig. 2). Mortality rates were higher amongst patients with hip fracture compared with those without, with a 5.0% (24/480) 30-day mortality, 12.5% (60/480) 120-day mortality, and 21.7% (105/480) 1-yr mortality rate (Fig. 3).

Amongst hip fracture participants, 69.8% (338/484) were female, with a mean age of 81.6 (sd 9.0) yr. In the non-hip fracture group, 47.6% were female, and the mean age was younger at 72.2 (sd 7.8) yr. Participants with hip fracture had lower BMI compared with the rest of the SNAP-3 cohort. In those with hip fracture, 53.3% (250/469) were of normal weight, and 9.2% (43/469) were underweight. In those without hip fracture, 28.2% (1846/6531) were of normal weight, and 1.5% were underweight (95/6531). The ASA physical status was higher in patients with hip fracture compared with those without: amongst those with hip fracture 70.2% (339/483) were ASA physical status 3 or 4, compared with 37% (2412/6516) of those without hip fracture. Of those undergoing surgery for hip fracture, 25.4% (123/484) did not have capacity to consent to this study, with 19.4% (94/484) having dementia. Two-thirds needed assistance with ADLs (help with instrumental ADLs 53.9% [229/425] and basic ADLs 10.8% [46/425]). Only 35.3% (150/425) were independent, despite 80.8% (391/484) being admitted from their own home. Polypharmacy was common with a prevalence of 58.8% (280/476). The prevalence of frailty was notably higher amongst patients with hip fracture; 58.2% (280/481) were living with frailty, compared with 16.7% (1089/6516) of those without hip fracture. Multimorbidity was also more common in patients with hip fracture, affecting 74.9% (346/462), compared with 62.2% (3632/5841) of those without hip fracture (Fig. 1).

Median postoperative length of stay was 13 (IQR 7–20) days for those with hip fracture. The incidence of postoperative delirium was 36.8% (178/484), and of other complications was 85.3% (413/484) (Fig. 2). Mortality rates were higher amongst patients with hip fracture compared with those without, with a 5.0% (24/480) 30-day mortality, 12.5% (60/480) 120-day mortality, and 21.7% (105/480) 1-yr mortality rate (Fig. 3).

These data highlight the heterogeneity of older surgical patients in clinically important subgroups, emphasising the importance of perioperative pathways tailored to the needs of patients, clinicians, and managers. Our key findings can be summarised as: (i) the older surgical population is distinctly heterogeneous in terms of characteristics, procedures performed, and outcomes; (ii) day surgery rates are relatively low; (iii) patients undergoing non elective surgery are more frail, older, and have worse outcomes; and (iv) the incidence of delirium is not evenly distributed across older surgical patients, with patients undergoing non elective surgery at particular risk. These findings are consistent with previous UK and international data demonstrating associations between older age or non elective surgery, and increased rates of frailty and adverse outcomes. Importantly the SNAP-3 data provide comparative, prospectively gathered granular information from a general cohort, in comparison with single specialty or administrative datasets. There are few data in the literature describing the older daycase surgery population beyond guidelines stating that age should not be an exclusion criterion. The hip fracture data are consistent with data from UK hip fracture databases, supporting the external validity of SNAP-3.37

specialty or administrative datasets. There are few data in the literature describing the older daycase surgery population beyond guidelines stating that age should not be an exclusion criterion. The hip fracture data are consistent with data from UK hip fracture databases, supporting the external validity of SNAP-3.37 In the short to medium term, the surgical population is likely to have greater absolute and relative numbers of older people, and levels of frailty within the general population are predicted to double by 2041.26 Findings from this and other analyses of the SNAP-3 cohort reinforce the need for continued investment in perioperative medicine services, which vary widely across hospitals, but have been shown to improve outcomes, shorten hospital stays, and enhance financial efficiency.27, 28, 29, 30 These subgroup data have utility in informing ongoing service development: the role of daycase pathways; differences between elective and non elective surgical populations; using data to support better-informed shared decision-making; and appropriate targeting of resources. Not all older patients are of high risk, thus routine screening for frailty, cognitive impairment, and multimorbidity is recommended to streamline access to appropriate care.31, 32, 33 Frailty incidence differs between elective and non elective populations; the proportion of patients with multimorbidity was similar. However, the SNAP-3 survey of publicly funded hospitals in the UK and Ireland in 2022 found that only half of preoperative assessment clinics screen ‘high-risk’ older patients for frailty, and just 40% screen for cognitive impairment, highlighting opportunities to improve perioperative screening.27 We have deliberately reported the cross-tabulated data on characteristics, outcomes, and anaesthetic techniques in full in the supporting information. The data presented here can be used to provide robust estimates of frailty and adverse outcomes in relevant groups to assist in efficient trial design or to explain effects of quality improvement or service changes.

ted data on characteristics, outcomes, and anaesthetic techniques in full in the supporting information. The data presented here can be used to provide robust estimates of frailty and adverse outcomes in relevant groups to assist in efficient trial design or to explain effects of quality improvement or service changes. We found that a third of all participants aged 60 yr and older had day surgery. Even when elective surgery was considered separately, the prevalence of day surgery was only 41%. Low rates of day surgery in this older cohort might be attributable to cautious decision-making by clinicians, concerned about lack of adequate social support at home.34 The impact of frailty and multimorbidity on postoperative outcomes might also encourage clinicians to plan inpatient surgery, sometimes overlooking the fact that overnight admission in itself is associated with complications in older adults.11,35 In particular, future work should focus on subgroups of older patients who could safely benefit from day surgery, and to establish evidence-based exclusion criteria.

icians to plan inpatient surgery, sometimes overlooking the fact that overnight admission in itself is associated with complications in older adults.11,35 In particular, future work should focus on subgroups of older patients who could safely benefit from day surgery, and to establish evidence-based exclusion criteria. Effective shared decision-making involves providing information that aligns with patient values and priorities. By examining the postoperative outcomes of specific patient subgroups, clinicians can offer more detailed and personalised insights into the risks associated with various types of surgeries, including elective, non elective, daycase, inpatient, orthopaedic, urology, colorectal, and hip fracture procedures. Our study group aims to further enhance these shared decision-making discussions by developing a quantitative risk-prediction tool for postoperative delirium, providing patients and clinicians with an even clearer understanding of the potential risks involved.

hopaedic, urology, colorectal, and hip fracture procedures. Our study group aims to further enhance these shared decision-making discussions by developing a quantitative risk-prediction tool for postoperative delirium, providing patients and clinicians with an even clearer understanding of the potential risks involved. The strengths and limitations of the SNAP-3 cohort study have been described.7 In summary, the primary strength of our study is its broad and representative cohort, captured from the older surgical population across the UK. Additional strengths include a low probability of selection bias, given the 80% hospital participation rate nationwide and good recruitment levels; a rigorous approach to identifying individuals with frailty, multimorbidity, and delirium; careful efforts to recruit those lacking capacity to consent; and low missing data in both initial and follow-up assessments. We believe the missing data are missing at random, though as with any observational study this is impossible to prove. Limitations include the possibility of not recruiting all eligible participants; potential underrepresentation of individuals without capacity; data that might not fully reflect weekend working patterns; and the presence of some missing data, though this remains low for an observational study (Supplementary Table S16). We have confidence that the cohort we report is representative and generalisable of the older surgical population in the UK based on comparable data from independent datasets such as the National Hip Fracture Database, National Emergency Laparotomy Audit, and NAP7. It is also likely that our findings will be relevant to similar high-income country healthcare settings globally, with similarities noted with recently published administrative data from Sweden examining older surgical patients.36

uch as the National Hip Fracture Database, National Emergency Laparotomy Audit, and NAP7. It is also likely that our findings will be relevant to similar high-income country healthcare settings globally, with similarities noted with recently published administrative data from Sweden examining older surgical patients.36 The study was not powered for subgroup analyses. We have deliberately avoided any inferential analyses or complex modelling because of the relatively small numbers. It is likely that the associations we describe are explained at least in part by selection biases before surgery; for instance, fitter patients might be more likely to be offered and accept elective surgery, and hence have better outcomes. We have explored these issues in our main analyses examining the associations between frailty, multimorbidity delirium, and postoperative outcomes. Given the smaller numbers of patients included in these subgroups, such analyses would have been of limited value here.

pt elective surgery, and hence have better outcomes. We have explored these issues in our main analyses examining the associations between frailty, multimorbidity delirium, and postoperative outcomes. Given the smaller numbers of patients included in these subgroups, such analyses would have been of limited value here. People with hip fracture and similar injuries sit at the intersection of those who are likely to be older, more often living with frailty and undergoing non elective and orthopaedic surgery. Numerically they clearly have an impact on the average characteristics and outcomes in each of those groups, given that they comprise around one-third of non elective surgery in those living with frailty, and around half of non elective orthopaedic surgery. They are a higher-risk group, and outcomes (calculable from data presented in the supporting information) for non-hip fracture, non elective orthopaedic older surgical patients are similar to those for non elective colorectal and urology patients.

iving with frailty, and around half of non elective orthopaedic surgery. They are a higher-risk group, and outcomes (calculable from data presented in the supporting information) for non-hip fracture, non elective orthopaedic older surgical patients are similar to those for non elective colorectal and urology patients. Finally, in the context of decision-making by patients and clinicians, we have no data on those who never had an operation. This is a notoriously difficult area to study as patients might not be offered an operation, or might make a choice not to pursue an operation at any point from the onset of symptoms or point of diagnosis through to the induction of anaesthesia. There is work ongoing in this area such as the no-laparotomy project from the National Emergency Laparotomy Audit, and the OSIRIS study into patient decision-making. SNAP-3 can offer insights into what happens after surgery, but cannot describe the alternatives.

int of diagnosis through to the induction of anaesthesia. There is work ongoing in this area such as the no-laparotomy project from the National Emergency Laparotomy Audit, and the OSIRIS study into patient decision-making. SNAP-3 can offer insights into what happens after surgery, but cannot describe the alternatives. This study describes the prevalence of frailty and the incidence of postoperative complications in important surgical subgroups with the aim of highlighting areas where perioperative services must be developed. Common themes amongst the seemingly disparate subgroups include the need for realistic choices underpinned by informed shared decision-making; consideration of specific needs of the oldest old, non elective surgical patients, in particular those after hip fracture; and the importance of delivering age-appropriate perioperative services including day surgery without regional variations. These findings should aid and encourage development of surgical services to provide patient-centred care from contemplation of surgery to full recovery.

This study describes the prevalence of frailty and the incidence of postoperative complications in important surgical subgroups with the aim of highlighting areas where perioperative services must be developed. Common themes amongst the seemingly disparate subgroups include the need for realistic choices underpinned by informed shared decision-making; consideration of specific needs of the oldest old, non elective surgical patients, in particular those after hip fracture; and the importance of delivering age-appropriate perioperative services including day surgery without regional variations. These findings should aid and encourage development of surgical services to provide patient-centred care from contemplation of surgery to full recovery.

Project initiation, guarantor, and grant holder: IKM Patient-facing document design and patient and public insights: BE, CJS, JSLP, IKM Ethical and regulatory approvals: CJS, KW Data collection tools and protocol design: CJS, JSLP, IKM, TP, AS Expertise in geriatric medicine: JSLP Expertise in anaesthesia and perioperative medicine: IKM, AS, TP Study implementation in the UK: CJS, KW, IKM, JSLP Site and UK-wide activity coordination: KW, CJS Analysis plan development: IKM, CJS Monitoring data collection: CJS Statistical expertise in study design and analysis methods: PM, HAB Data cleaning: CJS, IKM Data analysis: CJS, PM, HAB, IKM Review of final draft and contributions to revisions: all authors

This work was supported by the Frances and Augustus Newman Foundation and the Royal College of Anaesthetists. They had no input into the design, conduct or analysis of the study. HAB is supported by the 10.13039/501100000272National Institute for Health and Care Research ARC North Thames.

IKM is the Director of the Centre for Research and Improvement at the Royal College of Anaesthetists. IM, AS, and JP have received grant funding for clinical trials in perioperative care of older people. AS has received honoraria from Pharmacosmos UK outside of the submitted work.