Oncology Prognosis Tool

Mar 4, 2021 | Projects

The Oncology Prognosis Tool is a classifier used to determine the eligibility of patients for serious illness conversation. It was deployed at the Mount Sinai Oncology Outpatient unit in March 2021 and Inpatient unit in August 2022.

Challenges

  • Majority of patients(~ 80%) with advanced cancer want to stay informed about Prognosis 1
  • The most common prognostic approach is Clinician Prediction of Survival (CPS) 1
  • Clinicians intuition is often said to be optimistically biased2 which may result in
    • Aggressive Treatment
    • Being reluctant to communicate prognostic information to patients 3,4,5

Opportunity: Optimize confidence of clinicians’ prognostic judgment

Goal: Develop a classifier to determine the eligibility of patients for serious illness conversation

 

Introduction

  • Accurate prognostication of survival for patients with advance cancer is vital (1)
    • It informs many personal and clinical decisions
  • Prognosis is a process rather than a single event and patient’s prognosis may change over time based on: (3)
    • Treatment response
    • Development of complications
    • Competing co-morbidities
  • Prognostic accuracy varies by:(1)
    • patient population
    • Setting
    • Time frame of prediction
  • 10 major themes for the future prognostic research (1)
    • Enhancing prognostic accuracy
    • Improving reliability and reproducibility of prognosis
    • Identifying the appropriate prognostic tool for a given setting
    • Predicting the risks and benefits of cancer therapies
    • Predicting survival for pediatric populations
    • Translating prognostic knowledge into practice
    • Understanding the impact of prognostic uncertainty
    • Communicating prognosis
    • Clarifying outcomes associated with delivery of prognostic information
    • Standardizing prognosis terminology
  • Dependents of patient’s prognosis: (2,3)
    • Complex decisions regarding initiation, intensity, termination of palliative systemic therapies
    • Palliative procedures or surgery
    • Artificial nutrition or hydration
    • Hospice care
  • Clinician Prediction of survival (CPS) is the most common approach (22,23,24,25)
    • Temporal approach (how long)
    • Surprise questions (died in specific time frame? )
    • Probabilistic approach (probability of survival in specific time frame)
  • Early Palliative Care has been shown to improve health related quality of life (HRQOL),mood, and symptom score (34,35,36,37)
  • Early PC may increase patient satisfaction with care

Problem

  • Approximately 80 percent of patients with advanced cancer want to be informed of their:(1)
    • prognosis
    • expected treatment outcome
    • Advert effects
    •  Bodily changes in the last weeks to days of life
  • Patients hesitate to ask about prognosis directly and expect clinicians initiate such a conversation (1)
  • Caregivers have strong desire to be informed
  • Clinician intuition is often inaccurate (4)
    • Too optimistic may result in overly aggressive cancer treatment
    • Physicians are reluctant to communicate prognostic information to patients (5,6,7,9)
      • Lack of confidence in the accuracy of their prognostic judgment
      • Their judgments have been shown to be optimistically biased
        • Lack of historical evidence based prognostic information (10, 11)
      • barriers may inhibit physicians’ capacity or willing to communicate
        • Structural factors (5,10,17,18 )
          • Lack of time
          • Lack of financial incentives
        • Situational factors (10)
          • Clinical circumstances
            • Disease
              • acuity
              • Severity
              • Trajectory
          • Patient characteristics
          • Available clinical information
        • Emotional state of patients or patient emotional distress (19)
        • Irreducible uncertainty arising from the reliability, credibility and adequacy of all prognostic measures (20,21)
          • Individual differences in physicians’ tolerance pf ambiguity
  • The most common prognostic approach is: Clinician Prediction of Survival (CPS) —> Probabilistic (1)
  • Some of prognostic factors in oncology do not appear to have prognostic impact on advanced cancer patients in PC (34,39,40)
    • Tumor staging
    • Histologic grade
    • Genetics
  • There us a lack of studies that evaluated the prognosis of outpatients with advanced cancer who are receiving anti-neoplastic treatment and concomitantly undergoing PC (34)
    • Most of the studies describe a sample of patients with advanced disease , a low functional performance and a short life expectancy. (41)

Current Approaches and Solutions

  • Examples are including:(1)
    • Glasgow Prognostic Score (GPS) – common
    • Palliative Performance Scale (PPS) – common
    • Palliative Prognostic Score (PaP) – common
    • Palliative Prognostic Index (PPI) – common
    • Prognostic in  Palliative Care Study (PiPS)
    • Objective Prognostic Score (OPS)
  • Clinical Prediction models (CPMS) have been increasingly disseminated —> https://eprognosis.ucsf.edu/
    • A growing number of CPMs estimate short term (6-12 months) survival
      • Useful for acute critical EOL care decisions
        • Pursue palliative versus curative care (12, 13,14,15,16)
  • Predict toxicity of therapy in patients with advance cancer

Main Goal and Objectives

Proposed Solution

Target Audience

    • Family Medicine
    • Oncology

Proposed Trigger for Prediction

  • Realtime Prediction
    • Patients have the selected primary tumor site —> will be defined by Dr. Cardinale Smith
    • Have outpatient/ER visit number
  • Prediction on Demand
    • Physician put the request in Epic
    • Streaming CDS pipeline will receive it
    • Generate prediction and write it into a epic flowsheet in Epic

Patients and Methods

  • Inclusion Criteria
    • Age >= 21 yrs
    • Patients should be in the Cancer Registry
    • Primary Tumor Site should be in he selected primary tumor site —> with count >=200
    • Followup time in the CA registry >= 90 days  
  • Prognostic factors (26,27,28,29,30)
    • Deterioration in performance status (PS)
      • Palliative Performance scale
      • Karnofsky Performance status
      • ECOG performance status
    • Presence of distant metastasis
    • dyspnea at rest
    • Edema
    • Ascite
    • Malnutrition Assessment
    • Oral Intake
    • Level of consciousness
    • CAM Assessment
    • Cognitive Failure
    • Pain
    • Cancer anorexia-cachexia syndrome
      • reduced food intake
      • abnormal metabolism
      • Poor appetite
      • Nutrition impact symptoms (NIS)
        • Nausea
      • Weight loss
      • BMI
      • Changes in body composition
      • Sarcopenia
      • Inflammation
      • Testosterone  level
    • Cachexia staging score
    • Eastern Cooperative Oncology Group (ECOG)
    • SIRS
    • CRP
    • Albumin
    • WBC
    • Lymphocyte percentage
    • Phase Angle (PA)
    • Tell-tale signs (31,32)
      • Pulselessness of the radial artery
      • Hyperextension of the neck
      • Grunting of vocal cords
      • Cheyne-Stokes breathing
      • Death rattle
      • loss of facial muscle tone (drooping ofthe nasolabial folds) (33)
    • Age
    • Metabolism
    • Co-morbidities (1)
      • Rheumatoid Arthritis
      • Psoriasis
    •  Cancer Type
    • Planned Chemotherapy dose
      • standard
      • Reduced
    • Planned number of chemotherapy drugs
      • Poly >1
      • Mono = 1
    • Hemoglobin
    • Creatinine Clearance
      • < 34 ml/min
      • >= 34 ml/min
    • Hearing —> fair, poor,totally deaf, excellent or good
    • Number of fans in the past 6 months
    • Level of activity

Data Dictionary: ONC_DDV01232020.xlsx

Historical Cohort Population Structure: AA_COHORTV02102020.html

Presentation Slide Deck: AA_WG_long_version_V09032020.pptx (Long Version)

AA_WG_version_V10062020.pptx (Short Version)

Big Data Presentation Slide Deck, presented on November 9, 2020: AA_BigData_version_V11092020.pptx

  • Modeling approach
    • Multivariate cox regression (34)
      • Backward variable elimination
        • Discrimination properties were  evaluated
          • AUC score
      • Kolmogorov-Smirnov (K-S) goodness of fit —> distance between empirical distribution function and cumulative distribution function ——> measure the ability of prognostic tools
    • Barretos Prognostic Nomogram (BPN)
    • Calibration based on Hosmer-Lemeshow test

Performance Evaluation

  • Each prediction will be validated retrospectively and prospectively within the prediction window of 30 days, 60 days, and 90 days

References

  1. Hui D, Paiva CE, Fabbro EG Del, Steer C, Naberhuis J, Van M. Prognostication in advanced cancer : update and directions for future research. 2019.
  2. Glare PA, Sinclair CT (2008) Palliative medicine review: prognos- tication. J Palliat Med 11(1):84–103. https://doi.org/10.1089/jpm. 2008.9992
  3. Hui D (2015) Prognostication ofsurvival in patients with advanced cancer: predicting the unpredictable? Cancer Control 22(4):489– 497. https://doi.org/10.1177/107327481502200415
  4. Han PK, Dieckmann NF, Holt C,Gutheil C, Peters E (2016) Factors affecting physicians’ intentions to communicate personalized prog- nostic information to cancer patients at the end of life: an experi- mental vignette study. Med Decis Making 36(6):703–713. https:// doi.org/10.1177/0272989×16638321
  5. Hancock K, Clayton JM, Parker SM, et al. Truth-telling in dis- cussing prognosis in advanced life-limiting illnesses: a systematic review. Palliat Med. 2007;21(6):507–17. 3.
  6. Lamont EB, Christakis NA. Prognostic disclosure to patients with cancer near the end of life. Ann Intern Med. 2001;134(12): 1096–105. 5.
  7. Daugherty CK, Hlubocky FJ. What are terminally ill cancer patients told about their expected deaths? A study of cancer physi- cians’ self-reports of prognosis disclosure. J Clin Oncol. 2008; 26(36):5988–93.
  8. Cartwright C, Onwuteaka-Philipsen BD, Williams G, et al. Phy- sician discussions with terminally ill patients: a cross-national comparison. Palliat Med. 2007;21(4):295–303.
  9. Gattellari M, Voigt KJ, Butow PN, Tattersall MH.When the treat- ment goal is not cure: are cancer patients equipped to make informed decisions? J Clin Oncol. 2002;20(2):503–13. 8.
  10. Christakis NA, Lamont EB. Extent and determinants of error in physicians’ prognoses in terminally ill patients: prospective cohort study. West J Med. 2000;172(5):310–3.
  11. Glare P, Virik K, Jones M, et al. A systematic review of physi- cians’ survival predictions in terminally ill cancer patients. BMJ. 2003;327(7408):195–8.
  12. Teno JM, Harrell FE Jr, Knaus W, et al. Prediction of survival for older hospitalized patients: the HELP survival model. Hospital- ized Elderly Longitudinal Project. J Am Geriatr Soc. 2000;48(5 Suppl):S16–24. 14.
  13. Fischer SM, Gozansky WS, Sauaia A, Min SJ, Kutner JS, Kramer A. A practical tool to identify patients who may benefit from a palliative approach: the CARING criteria. J Pain Symptom Manage. 2006;31(4):285–92. 15.
  14. Inouye SK, Bogardus ST Jr, Vitagliano G, et al. Burden of ill- ness score for elderly persons: risk adjustment incorporating the cumulative impact of diseases, physiologic abnormalities, and functional impairments. Med Care. 2003;41(1):70–83.
  15. Walter LC, Brand RJ, Counsell SR, et al. Development and val- idation of a prognostic index for 1-year mortality in older adults after hospitalization. JAMA. 2001;285(23):2987–94.
  16. Han PK, Lee M, Reeve BB, et al. Development of a prognostic model for six-month mortality in older adults with declining health. J Pain Symptom Manage. 2012;43(3):527–39.
  17. Clayton JM, Hancock KM, Butow PN, et al. Clinical practice guidelines for communicating prognosis and end-of-life issues with adults in the advanced stages of a life-limiting illness, and their caregivers. Med J Aust. 2007;186(12 Suppl):S77, S79, S83–108.
  18. Lamont EB, Christakis NA. Physician factors in the timing of cancer patient referral to hospice palliative care. Cancer. 2002; 94(10):2733–7.
  19. Hallen SA, Hootsmans NA, Blaisdell L, Gutheil CM, Han PK. Physicians’ perceptions ofthe value ofprognostic models: the ben- efits and risks ofprognostic confidence. Health Expect. 2015;18(6): 2266–77. 22.
  20. Ellsberg D. Risk, ambiguity, and the Savage axioms. Quart J Econ. 1961;75:643–69.
  21. Camerer C, Weber M. Recent developments in modeling pref- erences: uncertainty and ambiguity. J Risk Uncertain. 1992;5: 325–70.
  22. Parkes CM (1972) Accuracy of predictions of survival in later stages ofcancer. Br Med J 2(5804):29–31
  23. White N, Reid F, Harris A, Harries P, Stone P (2016) A systematic review ofpredictions ofsurvival in palliative care: how accurate are clinicians and who are the experts? PLoS One 11(8):e0161407. https://doi.org/10.1371/journal.pone.0161407
  24. Amano K,Maeda I, Shimoyama S, Shinjo T, Shirayama H,Yamada T, Ono S, Yamamoto R, Yamamoto N, Shishido H, Shimizu M, Kawahara M, Aoki S, Demizu A, Goshima M, Goto K, Gyoda Y, Hashimoto K, Otomo S, Sekimoto M, Shibata T, Sugimoto Y, Morita T (2015) The accuracy of physicians’ clinical predictions of survival in patients with advanced cancer. J Pain Symptom Manag 50(2):139–146 e131. https://doi.org/10.1016/j. jpainsymman.2015.03.004
  25. Cheon S, Agarwal A, Popovic M, Milakovic M, Lam M, Fu W, DiGiovanni J, Lam H, Lechner B, Pulenzas N, Chow R, Chow E (2016) The accuracy of clinicians’ predictions of survival in ad- vanced cancer: a review. Ann Palliat Med 5(1):22–29. https://doi. org/10.3978/j.issn.2224-5820.2015.08.04
  26. Maltoni M, Caraceni A, Brunelli C, Broeckaert B, Christakis N, Eychmueller S, Glare P, Nabal M, Vigano A, Larkin P, De Conno F, Hanks G, Kaasa S (2005) Prognostic factors in advanced cancer patients: evidence-based clinical recommendations–a study by the Steering Committee of the European Association for Palliative Care. J Clin Oncol 23(25):6240–6248. https://doi.org/10.1200/ JCO.2005.06.866
  27. Trajkovic-Vidakovic M, de Graeff A, Voest EE, Teunissen SC (2012) Symptoms tell it all: a systematic review of the value of symptom assessment to predict survival in advanced cancer pa- tients. Crit Rev Oncol Hematol 84(1):130–148. https://doi.org/10. 1016/j.critrevonc.2012.02.011
  28. Lasheen W, Walsh D (2010) The cancer anorexia-cachexia syn- drome: myth or reality? Support Care Cancer 18(2):265–272. https://doi.org/10.1007/s00520-009-0772-6
  29. Dolan RD, McSorley ST, Horgan PG, Laird B, McMillan DC (2017) The role of the systemic inflammatory response in predicting outcomes in patients with advanced inoperable cancer: systematic review and meta-analysis. Crit Rev Oncol Hematol 116: 134–146. https://doi.org/10.1016/j.critrevonc.2017.06.002
  30. Hui D, Bansal S, MorgadoM, Dev R, Chisholm G, Bruera E (2014) Phase angle for prognostication of survival in patients with ad- vanced cancer: preliminary findings. Cancer 120(14):2207–2214. https://doi.org/10.1002/cncr.28624
  31. Hui D, dos Santos R, Chisholm G, Bansal S, Silva TB, Kilgore K, Crovador CS, Yu X, Swartz MD, Perez-Cruz PE, Leite Rde A, Nascimento MS, Reddy S, Seriaco F, Yennu S, Paiva CE, Dev R, Hall S, Fajardo J, Bruera E (2014) Clinical signs of impending death in cancer patients. Oncologist 19(6):681–687. https://doi. org/10.1634/theoncologist.2013-0457
  32. Hui D, Dos Santos R, Chisholm G, Bansal S, Souza Crovador C, Bruera E (2015) Bedside clinical signs associated with impending death in patients with advanced cancer: preliminary findings of a prospective, longitudinal cohort study. Cancer 121(6):960–967. https://doi.org/10.1002/cncr.29048
  33. Hui D, Hess K, dos Santos R, Chisholm G, Bruera E (2015) A diagnostic model for impending death in cancer patients: prelimi- nary report. Cancer 121(21):3914–3921. https://doi.org/10.1002/ cncr.29602
  34. Paiva, Carlos Eduardo,Paiva, Bianca Sakamoto Ribeiro,de Paula Pântano, Naitielle,Preto, Daniel D’Almeida,de Oliveira, Cleyton Zanardo,Yennurajalingam, Sriram,Hui, David,Bruera, Eduardo (2018)Development and validation of a prognostic nomogram for ambulatory patients with advanced cancer J Cancer Medicine 7(7):3003–3010
  35. Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic non- small- cell lung cancer. N Engl J Med. 2010;363:733‐742.
  36. Bakitas M, Lyons KD, Hegel MT, et al. Effects of a palliative care intervention on clinical outcomes in patients with advanced can- cer: the Project ENABLE II randomized controlled trial. JAMA. 2009;302:741‐749.
  37. Follwell M, Burman D, Le LW, et al. Phase II study of an outpa- tient palliative care intervention in patients with metastatic can- cer. J Clin Oncol. 2009;27:206‐213.
  38. McDonald J, Swami N, Hannon B, et al. Impact of early palliative care on caregivers of patients with advanced cancer: cluster ran- domised trial. Ann Oncol. 2017;28:163‐168
  39. Krishnan M, Temel JS, Wright AA, Bernacki R, Selvaggi K, Balboni T. Predicting life expectancy in patients with advanced incurable cancer: a review. J Support Oncol. 2013;11:68‐74.
  40. Glare P. Clinical predictors of survival in advanced cancer. J Support Oncol. 2005;3:331‐339.
  41. Simmons CPL, McMillan DC, McWilliams K, et al. Prognostic tools in patients with advanced cancer: a systematic review. J Pain Symptom Manage. 2017;53:962‐970.e10.

Diagnosis Grouping Lookup: (source: Mark Liu)

ICD-10 Code Mapping_updatedML.xlsx