Imminence of death in hospital patients: evolution of a study - PowerPoint Presentation

1. Imminence of death in hospital patients: evolution of a studyDavid Clark @dumfriesshireWellcome Trust InvestigatorUniversity of Glasgow

2. A selection of initiatives on hospital end of life careHospice Friendly Hospitals (Ireland)Living and Dying Well (Scotland)The ‘Transform’ programme (England)PATCH (Scotland)Building on the Best (England+Scotland)

3. Royal College of Physicians audit (2016)Major gaps in documenting basic aspects of care in patients whose death could reasonably be anticipatedThese included patients or loved ones discussing their own concerns and choices; spiritual care, clear plans around eating, drinking, artificial nutrition, and hydration; and control of common symptomsOnly a third of hospitals had face to face specialist palliative care seven days a week. Many patients were unable to see specialist doctors or nursesSeven key recommendations – SPC provision, education/training, audit, designated board member, assessment of imminently dying, four hourly monitoring of symptoms, pastoral and spiritual care provision

4. Quality of care in hospitals (CQC March 2017)292 hospitals in England - variation both in the quality of care between hospitals and between individual core services within the same hospital. While the majority of hospital services are delivering good quality care and looking after patients well, inspections have also uncovered pockets of poor care even in good hospitals..End of life care services: rated outstanding (7%), good (56%) requiring improvement (33%) inadequate (4%) The NHS stands on a burning platform - the model of acute care that worked well when the NHS was established is no longer capable of delivering the care that today’s population needs

5. The original study of imminence of death among Scottish hospital patientsPrevalent cohort study on one day Population based, not a sampleAll acute hospitals in ScotlandLinking hospital records to death registrationsFirst study of its kind in the world

6. In hospital – 2010 Scotland Census of 10,743 inpatients in 25 Scottish teaching and general hospitals on 31 March 2010 3,098 (28.8%) patients died during the one year follow up period: 2.9% by 7 days, 8.9% by 30 days, 16.0% by 3 months, 21.2% by 6 months, 25.5% by 9 months and 28.8% by 12 months.Almost one in ten patients (9.3%) patients died during the index admission

7.

8. Follow up in hospital – Scotland 201310,595 patients were in hospital on 10 April 201329.5% of the 2013 cohort had died one year later 8.0% died on the index admissionPrevious result confirmed … no significant differencesKaplan Meier plot comparing survival curves of 2010 and 2013 cohorts

9.  Patients in Hospital on 31/03/2010 who died within 12 monthsPatients in Hospital on 10/04/2013 who died within 12 months No.%No.%Gender    Male147730.4%151031.0%Female161627.5%161628.3%Age    Under 6039413.1%33311.8%60-6419122.6%16621.4%65-6929929.3%30629.6%70-7435530.0%39634.7%75-7949033.0%49235.1%80-8455939.2%58939.0%85 and over80545.6%84444.3%Deprivation: SIMD (2012)    1 (Most)87129.7%87229.6%271930.2%74030.0%356927.0%53628.3%450428.1%50629.5%5 (Least) 42430.8%46730.8%Missing69.8%59.1%Specialty    Surgical68517.3%63017.7%Medical230435.0%240335.1%Other10451.5%9352.0%Total3093 (adjusted)28.8%312629.5%

10. ImplicationsTwo purposes in repeating the analysis for 2010 / 2013: verification, change over timeIt would appear that Scottish hospitals are in a reasonably settled pattern of containing a substantial proportion of inpatients that are in the last year of lifeThere was no significant difference in this between the two cohorts

11. ‘Impact’ of first studyGoogle Scholar citations = 67 Widely known, often featured in clinical presentationsContinues to be mentioned on social media discussionsAltmetrics score = 286Outputs from Palliative Medicine #3 of 1,125

12.

13. Latest Scottish study (Chris Isles et al, BMJ Open - under review following revision)Incident cohort studyAll (10,477) emergency hospital admissions to medicine 18-25 March 2015 in Scotland22.4% died by 12 months. 5.8% died during the index admissionMortality five times higher for patients aged 85 years cf under 60 years 41.9% vs 7.9%). Men had higher mortality than women (24.1% v 20.8%).  Cancer patients had higher mortality rate at one year (57.6% v 17.6%). 68% of the deaths occurred in hospital48% of all patients had been admitted as an emergency to hospital in the year before the census admission

14. Isles et al forthcoming - Characteristics of cohort and associated one-year mortality ratesNo. of Admissions%Deaths within 7 days%Deaths within 30 days%Deaths within 3 months%Deaths within 6 months%Deaths within 9 months%Deaths within 1 year%SexFemale5,46352.1%1482.7%3356.1%66912.2%85915.7%1,03218.9%1,13820.8%Male5,01447.9%1312.6%3717.4%74514.9%93718.7%1,11322.2%1,20824.1%SIMDSIMD1 (most)3,09229.5%782.5%2046.6%37212.0%47315.3%56418.2%62120.1%SIMD22,47823.7%662.7%1556.3%34213.8%43617.6%51920.9%57423.2%SIMD31,84417.6%472.5%1196.5%25513.8%30916.8%37420.3%41122.3%SIMD41,66715.9%432.6%1217.3%23614.2%30518.3%37022.2%39723.8%SIMD5 (least)1,39613.3%392.8%946.7%19313.8%25718.4%30321.7%34324.6%Age group0-593,91537.4%270.7%802.0%1794.6%2275.8%2777.1%3097.9%60-647106.8%121.7%354.9%7911.1%10414.6%12517.6%13218.6%65-699268.8%272.9%646.9%13314.4%17218.6%20722.4%23425.3%70-741,06610.2%393.7%989.2%19218.0%22921.5%27025.3%29727.9%75-791,17511.2%272.3%897.6%20317.3%26322.4%31526.8%35129.9%80-841,21311.6%534.4%12510.3%24320.0%32026.4%37631.0%40633.5%85+1,47214.0%886.0%20213.7%36925.1%46531.6%56038.0%61741.9%Cancer diagnosisNo9,21487.9%2022.2%4595.0%8969.7%1,16312.6%1,44415.7%1,61817.6%Yes1,26312.1%715.6%23418.5%50239.7%61748.9%68654.3%72857.6%Total10,477100.0%2732.6%6936.6%1,39813.3%1,78017.0%2,13020.3%2,34622.4%

15. Incident cohort study- mortality after 12 months

16. To identify the proportion of a cohort of NZ public hospital inpatients dying within 12 months of a given census date.To identify associations between likelihood of death and key socio-demographic, diagnostic and service related factors.To compare results with, and extend findings of Scottish study undertaken for the same time period and census date.New Zealand – Prof Merryn Gott

17. MethodsData provided for all publicly-funded hospitalisations on 10 April 2013:demographic information (age, gender, deprivation of area of residence, prioritised self-identified ethnicity)hospitalisation (admission date and type, discharge date and type, length of stay, specialty and diagnosis)mortality information

18. Findings6,074 publicly-funded patients stayed overnight in NZ hospitals on the census date: 54% female; 46% male. 42% of patients were aged >60 years (17% of whom >80 years)878 (14.5%) patients had died during the 12-months following the census date: 1.6% by 7 days, 10.9% by 6 months, and 14.5% by 12 months135 patients (2.1%) died during the index stay - this accounted for 14.8% of all deaths within the 12-month follow-up period.

19. After 12 months (NZ)14.5% of patients had died 2.1% during the index stayOlder age and deprivation were predictors of dying

20. Denmark – Dr Lene JarlbaekIndex-date: 10th April 2013 (as for NZ and Scotland)Included were patients in all somatic, public hospitalsExcluded were in-patients in obstetric departments, and in-patients in psychiatric hospitals. 13,412 patients were in hospital or were admitted on the 10th April 2013, and stayed overnight (NB differences with Scotland and NZ)

21. Resultsdeadalive22% had died within 1 year after the index-date5.5% died on the index epsisodeHalf of the deceased died less than 60 days after the index-date

22. Hospitals - patients, proportions of deceased and age of the patients

23. Other potential studiesInterest from Hong KongThird study planned for Scotland using census date of 6 April 2016Still no version of the study from England, or Wales …

24. Further developmentsDr Marian Krawczyk, new Lord Kelvin Adam Smith FellowSystematic review and new ethnographic research in selected hospitals

25. Ethnographies of end of life in hospitals

26. Lifting the lid: ethnographic studies of end of life care in hospitals – systematic review Marian Krawcyzk et alBooks: 23 Book chapters: 10Journal articles: 150+Dissertations: 46Little known about this workKnowledge transfer problem

27. References Clark, D., Armstrong, M., Allan, A., Graham, F., Carnon, A., and Isles, C. Imminence of death among a national cohort of hospital inpatients. Palliative Medicine, 2014, 28(6): 474-479Clark, D Schofield, L Graham, F Gott, M Isles, C Jarlbaek, L (2016) Letter: Likelihood of death within one year among a national cohort of hospital inpatients in Scotland. Journal of Pain and Symptom Management. 52(2): e2-e4. Gott M, Broad J, Zhang X, Jarlbaek L, Clark D. Likelihood of death among hospital inpatients in New Zealand: prevalent cohort study. BMJ Open. 2017;7(12):e016880.Lene Jarlbaek, Helle Timm, Merryn Gott, David Clark, Imminence of death in Danish hospitals – a population-based study. In preparation for Scandinavian Journal of Public Health.

28. Glasgow End of Life StudiesGroupglasgow.ac.uk/endoflifestudies@EndofLifeStudyDavid Clark@dumfriesshire