Vol.

-

No.

- -

2015

Journal of Pain and Symptom Management

1

Original Article

Timing of POLST Form Completion by Cause of Death Dana M. Zive, MPH, Erik K. Fromme, MD, MCR, Terri A. Schmidt, MD, MS, Jennifer N.B. Cook, GCPH, and Susan W. Tolle, MD Center for Policy and Research in Emergency Medicine (D.M.Z.); Division of Hematology and Medical Oncology (E.K.F.); Department of Emergency Medicine (T.A.S., J.N.B.C.); Center for Ethics in Health Care (S.W.T.), Division of General Internal Medicine and Geriatrics, Oregon Health & Science University, Portland, Oregon, USA

Abstract Context. The physician orders for life-sustaining treatment (POLST) paradigm allows health care professionals to document the treatment preferences of patients with advanced illness or frailty as portable and actionable medical orders. National standards encourage offering POLST orders to patients for whom clinicians would not be surprised if they died in the next year. Objectives. To determine the influence of cause of death on the timing of POLST form completion and on changes to POLST orders as death approaches. Methods. This was a cohort study of 18,285 Oregon POLST Registry decedents who died in 2010e2011 matched to Oregon death certificates. Results. The median interval between POLST completion and death was 6.4 weeks. Those dying of cancer had forms completed nearer death (median 5.1 weeks) than those with organ failure (10.6 weeks) or dementia (14.5 weeks; P < 0.001). More than 90% of final POLST forms indicated orders for no resuscitation and 65.1% listed orders for comfort measures only. Eleven percent of the sample had multiple registered forms during the two years preceding their death, with the form completed nearest to death more likely than earlier forms to have orders for no resuscitation and comfort measures only, although some later forms did have orders for more treatment. Conclusion. More than half of POLST forms were completed in the final two months of life. Cause of death influenced when POLST forms were completed. POLST forms changed in the two years preceding death, more frequently recording fewer life-sustaining treatment orders than the earlier form(s). J Pain Symptom Manage 2015;-:-e-. Ó 2015 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved. Key Words POLST, end-of-life care, illness trajectories, timing of completion

Introduction Waiting too long to engage in advance care planning may result in late hospice referral and, for some, unwanted transitions in the final weeks of life. In 2011, the median national hospice stay was 19.1 days, with 35.8% of enrollees dying within one week of enrollment.1 Although rates of inhospital death are dropping, Teno et al.2 report that transitions during the last 90 days of life have increased from 2.1 in 2000 to 3.1 transitions in 2009. Some of these Address correspondence to: Dana M. Zive, MPH, Center for Policy and Research in Emergency Medicine, Oregon Health & Science University, CDW-EM, 3181 SW Sam Jackson Park Road, Portland, OR 97239, USA. E-mail: [email protected] Ó 2015 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved.

transitions were undoubtedly burdensome and/or unwanted. If called, most EMS protocols require that patients receive life-sustaining treatment and transport to a hospital unless EMS responders see written medical orders to the contrary; thus, having medical orders can help patients avoid unwanted treatment or transitions near the end of life. Physician orders for life-sustaining treatment (POLST) forms are medical orders completed by a health professional based on the patient’s preferences

Accepted for publication: June 1, 2015.

0885-3924/$ - see front matter http://dx.doi.org/10.1016/j.jpainsymman.2015.06.004

2

Zive et al.

regarding cardiopulmonary resuscitation in the case of cardiopulmonary arrest (do not resuscitate [DNR] or attempt resuscitation [cardiopulmonary resuscitation]), scope of treatment when not in cardiac arrest (comfort measures only [CMO], limited interventions, or full treatment), and artificial nutrition (long-term use, defined trial period, or no artificial nutrition by tube; see http://tinyurl.com/klrdob2 for a copy of the 2014 Oregon POLST form). Hickman et al.3,4 have shown that POLST reduces unwanted transitions. Tolle et al.5 reported a 5% inhospital death rate in a prospective study of nursing home residents with POLST orders for CMO. There is an association between POLST scope of treatment orders and location of death,6 as well as research indicating that POLST forms support patients in receiving the treatments documented on their POLST forms and avoiding those treatments they do not want.6,7 When is the optimal time to complete a POLST form? Although patients have the right to refuse treatments that they do not want, concerns have been raised that documents could be completed too early for patients to have a clear idea of what their preferences will be when they are more ill and nearer to death.8e10 Tuohey and Hodges11 provided a broader clinical context for the apprehension expressed by some writers that advance directives and POLST documents will result in limiting treatment if they are completed too soon, at a time when patients might live significantly longer if they did receive lifesustaining treatment. Limiting POLST use to patients who are identified as ‘‘terminally ill’’ potentially creates challenges,12 and many studies document physicians’ struggles with prognostic accuracy.13e15 Given challenges with prognostication, patient denial of terminal illness, and concerns with appropriate timing of advance care planning, the National POLST Paradigm Task Force recommends that clinicians have goals of care conversations and offer a POLST to patients whom they would not be surprised if they died in the upcoming year.16 The goal of this study was to examine the timing of POLST completion in Oregon and changes in POLST orders over time. Specifically, the objective was to determine the influence of the cause of death on the timing of POLST form completion as death approaches.

Vol.

-

No.

- -

2015

Study Setting The study included data from the Oregon POLST Registry, a statewide electronic registry of POLST forms. The completion of a POLST form by a patient is always voluntary, but, if one is completed in Oregon, the signing health care professional or their designee is mandated to submit the form to the Registry unless a patient specifically opts out. The Registry receives forms from all counties in the state and from all care settings, including long-term care, health systems, hospitals, clinics, hospice, home-based care, and individuals. Death certificates for 2010 and 2011 from the Oregon Center for Health Statistics were matched to decedent POLST forms in the Oregon POLST Registry. The primary outcome measured was the timing of POLST form completion in relation to death by cause of death. A similar sample was used in a prior analysis focused on location of death.6

Population The study included POLST forms for the 18,285 registrants in the Oregon POLST Registry for Oregonians who died of natural causes in 2010 or 2011. For each decedent, a two-year interval before date of death was calculated, and death records were matched to any registrant with a form submitted to the Registry within that interval. Only deaths by natural causes were included for the match, excluding suicide, homicide, accident, trauma, or undetermined/pending causes of death because most patients with sudden, unexpected, and traumatic deaths are usually not candidates to have POLST orders. Forms that had been signed within the two years before each registrant’s death were included. A subset of 2004 registrants (11%) was identified as having more than one registered POLST form in the two years before death that reflected different treatment preferences in at least one section as compared with an earlier form.

Variables Variables from death records from the Oregon Center for Health Statistics included dates of birth and death, race, ethnicity, educational attainment, decedent zip code (coded to rural or urban), sex, age at death, and primary cause of death. Variables from the Oregon POLST Registry included form orders for all POLST forms signed within two years of each decedent’s date of death, demographic data, and date each form was signed. Please see Fig. 1 for the POLST form in use during this study.

Methods This cohort study was reviewed by the Oregon State Public Health Institutional Review Board as well as the Oregon Health & Science University Institutional Review Board and was deemed exempt as all subjects are known decedents.

Statistical Analysis Descriptive univariate statistics were used to characterize the primary sample, using chi-square to determine significant differences between groups. Differences at the 0.05 level and below were considered

Vol.

-

No.

- -

2015

Timing of POLST Completion

3

Fig. 1. 2011 Oregon POLST form. POLST ¼ physician orders for life-sustaining treatment.

statistically significant. For continuous variables, both means and medians were assessed, including SDs for the former and interquartile ranges for the latter. Analyses were conducted using IBM SPSS Statistics, version 21 (IBM Corp., Armonk, NY).

Results Sample characteristics are summarized in Table 1. The registered decedents had a mean age of 79.1 years

(range: 0.01e106.53) at the time their final POLST form was signed. Forty-one percent died in 2010 and 58.9% in 2011. More than half were female (54.5%); 57.5% had urban residences at death (based on zip code); and 40.7% had completed at least some college. The median interval between final POLST form completion and death was 6.4 weeks (interquartile range: 19.9). In this two-year sampling time frame, 89% of earliest registered POLST forms, and 91.4% of final POLST forms were completed in the last year of life.

4

Zive et al.

Table 1 2010e2011 Oregon POLST Registry Decedent Characteristics Primary Sample (N ¼ 18,285) Female Mean age at signing of last POLST form (SD), years Location of residence Rural Ethnicity and racea White Hispanic or Latino African American Asian American Indian/Alaska Native Other or unknown More than one listed Hawaiian/Pacific Islander Education Less than high school HS/GED Some college Associates or bachelor’s degree Master’s, doctorate, or professional degree Unknown Causes of death Cancer Other natural causes Organ failureb Dementia

Vol.

-

No.

- -

2015

Table 2 POLST Orders on Final Registered Form Before Deatha

n (%) Sample or Subset 9965 (54.5) 79.1 (13.1) 7772 (42.5) 17,646 251 192 184 105 77 64 17

(96.5) (1.4) (1.1) (1.0) (0.6) (0.4) (0.4) (0.1)

3201 7423 3315 2936 1184 226

(17.5) (40.6) (18.1) (16.1) (6.5) (1.2)

6157 5128 4760 2240

(33.7) (28.0) (26.0) (12.3)

POLST ¼ physician orders for life-sustaining treatment; HS ¼ high school; GED ¼ general education development. a Only 4.5% of Oregonians who died in 2011 or 2012 were other than Caucasian. b It includes heart diseases, liver disease, chronic lower respiratory disease, and kidney disease.

Table 2 outlines the POLST orders on the final registered POLST form before death. The most common orders were durable DNR (92.3%), CMO (65.1%), and no artificial nutrition by tube (78.9%). The most common combination of POLST orders at the time of death was DNR and CMO (64.7%). Registrants who were nearer death at the time a POLST form was completed tended to have orders to for fewer life-sustaining treatments (Table 3). For example, registrants with orders for DNR and CMO had a POLST completed a median of 4.3 weeks before death. In comparison, for those with orders for attempt resuscitation and full treatment, the POLST was completed a median of 17.7 weeks before death. Of note, 47.8% of CMO orders were written for registrants who were in their final four weeks of life. We examined the relationship between cause of death and time interval from earliest registered POLST form to death. Table 3 outlines the median times to death for all decedents and orders, including order combination groups, as well as for three illness trajectories created from cause of death: Alzheimer’s disease or other causes of dementia; organ failure (including chronic lung disease, kidney disease, heart diseases, and liver disease); and cancer. Fig. 2

Final Registered FormdPrimary Sample, n (%)

All POLST forms All registered decedents’ forms 18,285 Resuscitation Do not resuscitate 16,869 (92.3%) Attempt resuscitation 1416 (7.7%) Scope of treatment Comfort measures only 11,896 (65.1%) Limited additional interventions 5033 (27.5%) Full treatment 1221 (6.7%) Artificial nutrition by tube No artificial nutrition by tube 14,429 (78.9%) Defined trial period of artificial nutrition 2695 (14.7%) by tube Long-term artificial nutrition by tube 617 (3.4%) Resuscitation and scope of treatment combinations DNR and comfort measures only 11,836 (64.7%) DNR and limited additional interventions 4640 (25.4%) DNR and full treatment 298 (1.6%) CPR and comfort measures only 60 (0.3%) CPR and limited additional interventions 393 (2.1%) CPR and full treatment 923 (5.0%) POLST ¼ physician orders for life-sustaining treatment; DNR ¼ do not resuscitate; CPR ¼ cardiopulmonary resuscitation. a All forms in the Registry are required to have an order regarding resuscitation. For scope of treatment, 135 forms did not have an order. For artificial nutrition by tube, 544 did not have an order. These ‘‘not filled out’’ sections are excluded from the table, so not all sections will equal 100%.

illustrates the differences in the interval between POLST form completion to death between the three illness trajectory groups, which were statistically significant (P < 0.001). Fig. 3 shows the changes in POLST orders as death approaches in the subset of 2004 registrants (11%) with more than one form in the Registry in the two years before their death. Of this subset, 1850 had two forms, 144 had three forms, and 10 had four forms. For each POLST order section, earlier POLST forms more frequently included orders for more life-sustaining treatment than final POLST forms. Thirty-two percent of registrants had orders to attempt resuscitation on their earlier form, whereas only 7.3% had such orders on their final form. Thirteen percent of registrants had orders for CMO on their earlier form, whereas 69.4% had such orders on their final form. Finally, 43.8% of registrants had orders for no artificial nutrition by tube on their earlier form, whereas 78.7% had such orders on their final form. Of the 2004 registrants with more than one form, 1567 (78.2%) had orders for fewer life-sustaining treatments on a subsequent form. Although the numbers were smaller for those having orders for more treatments nearer death, POLST form orders changed in both directions. Overall, 437 registrants with more than one form requested some increase

Vol.

-

No.

- -

2015

Timing of POLST Completion

5

Table 3 Interval in Weeks Between Final Registered Form Completion and Deatha Median Interval from Final Form to Death (N ¼ 18,285) Sample or Subset All POLST forms All registered decedents’ forms Cause of death/illness trajectory Alzheimer’s disease/dementia Organ failure Cancer Resuscitation Attempt resuscitation Do not resuscitate Scope of treatment Full treatment Limited additional interventions Comfort measures only Artificial nutrition by tube Defined trial period of artificial nutrition by tube Long-term artificial nutrition by tube No artificial nutrition by tube Resuscitation and scope of treatment combinations DNR and full treatment DNR and limited additional interventions DNR and comfort measures only CPR and comfort measures only CPR and full treatment CPR and limited additional interventions

Weeks (Interquartile Range) 6.4 (19.9)

P-Value n/a

14.5 (36.96) 10.6 (29.96) 5.1 (12.57)

Timing of POLST Form Completion by Cause of Death.

The physician orders for life-sustaining treatment (POLST) paradigm allows health care professionals to document the treatment preferences of patients...
2MB Sizes 2 Downloads 9 Views