Summary of included articles
| Reference | Conditions | Origin | Quality | Design | Patients included | Measures used | Results | Summary | Pros + and cons − |
|---|---|---|---|---|---|---|---|---|---|
| Allen et al16 2008 | HF | USA | Medium | Cross-sectional interviewer-administered questionnaire in a single centre outpatient heart-failure service | 122 sequentially recruited participants with HF (NYHAI-IV) Mean age 61 (IQR 53–74) 62% male 47% African-American |
| Median self-estimated life expectancy was 13 years (IQR 8–21; range 1–54 years) Median model-predicted life expectancy was 10 years (IQR 7.2–13.3; range 2.0–25 years) 66% of patients overestimated their survival compared with the model by 30% or more The median overestimate was 40% 29% of patients died within 3 years | Self-estimated-life expectancy was on average significantly greater than that predicted by a validated model Younger age, greater disease severity and measures of less depression were independently associated with overestimation of survival | + Efforts made to improve and check patient understanding of question − 26 of 148 enrolled participants felt unable/unwilling to estimate survival − Only 35 of 122 patients were followed up until their death − Only 9 of 122 patients had NYHA IV HF − No index group without chronic disease was included |
| Fried et al15 2003 | COPD HF | USA | Medium | Cross-sectional interview survey administered to patients registered at community practices and outpatient clinics of two hospitals, and inpatients of three hospitals. Same patient group as Fried et al 2006 | 135 patients with COPD or HF, aged 60 and older, meeting criteria for limited life expectancy and requiring assistance with daily living COPD—79 patients Mean age 72 (SD 7) 51% Male 92% White HF—56 patients Mean age 75 (SD 8) 70% Male 88% White | Patients and clinicians were asked how long they thought the patient would live and answered using multichoice options ranging from <1 month to >10 years | Only 9 of 135 patients expected to live <1 year, but 38 patients died over this period. 58 of 79 patients who responded to being asked to estimate their own life expectancy expected to live 2 years or more Of the 65 available patient–clinician pairs who both responded, 34 agreed the prognosis was 2 years or more, 9 agreed the prognosis was 2 years or less, 7 clinicians thought the patient would live 2 years or more when the patient did not expect to live this long and 15 patients expected to live 2 years or more when their clinician was less optimistic Kappa was 0.22 suggesting very poor agreement | Patient expectations of 1 year mortality are higher than observed. Agreement between patients and their clinicians about likely prognosis is poor. | − 56 of 135 patients were unable or unwilling to estimate their life expectancy − No index group without chronic disease was included |
| Fried et al14 2006 | COPD HF | USA | Medium | Serial interview survey administered to patients registered at community practices and outpatient clinics of two hospitals, and inpatients of three hospitals. Same patient group as Fried et al 2003 | 135 patients with COPD or HF, aged 60 and older, meeting criteria for limited life expectancy and requiring assistance with daily living COPD—79 patients Mean age 72 (SD 7) 51% Male 92% White HF—56 patients Mean age 75 (SD 8) 70% Male 88% White | Patients were asked how long they thought the patient would live and answered using multichoice options ranging from <1 month to >10 years | 9 of 59 patients who responded expected to live <1 year at their first interview. Of 59, 5 expected to live <1 year at their final interview 38 of 135 patients died over this period | Patient expectations of 1 year mortality are higher than observed The majority of patients (those who were alive and dead at the end of the year-long study) made no adjustment to their self-estimated life expectancy | − 56 of 135 patients were unable or unwilling to estimate their life expectancy − No index group without chronic disease was included |
| Kraai et al17 2013 | HF | The Netherlands | Low | Cross-sectional questionnaire administered in outpatient setting in one HF clinic. Subcomponent of time trade-off study | 100 patients with HF (NYHA I–IV) all over 50 years of age. Mean age 70 (SD 9.4) 71% male | Visual Analogue Scale from 50 to 100 years of age; patients were asked to indicate the most accurate estimation of their life expectancy | Mean life expectancy indicated by patients was 82 (SD 8.6) years. No difference in self-estimated life expectancy was found between patients unwilling vs willing to trade time | Self-estimated life expectancy probably exceeds likely outcomes, but no comparator data was available. Despite patients with more advanced or symptomatic HF being more willing to trade time, no difference was found between the groups in terms of expected longevity | − No comparator prediction or measurement of survival used −Only 2 of 100 patients had NYHA IV HF − No index group without chronic disease was included |
| Shah et al18 2006 | HF COPD CKD | UK | Low | Cross-sectional interviewer-administered questionnaire in outpatient and inpatient settings at one acute NHS Trust and a neighbouring hospice | 20 patients in total meeting criteria for limited life expectancy:6 HF (NYHA III/IV) 9 COPD 5 CKD Median age 72 50% male 85% white | Patients and physicians chose one of seven short prognosis statements that most accurately predicted how their illness might affect their life expectancy | 7 of 20 (35%) patients estimated their prognosis to be <1 year 13/17 physicians (76%) estimated their patient's prognosis to be < 1 year | Exploratory study, no firm conclusions available | − Very small numbers − Sample poorly representative of a general outpatient population − No index group without chronic disease was included |
| Stewart et al19 2010 | HF | USA | Low | Cross-sectional written questionnaire with inpatients and outpatients from two HF centres. Subcomponent of time trade-off study. | 105 patients with LVEF <35% and symptomatic HF Mean age 58 (SD 13) 70% male | Methodology for collecting self-estimated life expectancy not described | 65% thought they would live more than 10 years and 34% believed they would be alive for at least 20 years. Patients willing to trade more time expected shorter survival than those unwilling to trade time. 46% of the patients willing to trade away at least 12 months anticipated that they would not survive 5 years. No difference was found in self-estimated survival between inpatients and outpatients (data not provided) | Self-estimated life expectancy probably exceeds likely outcomes, but no comparator data was available. Willingness to trade time is associated with shorter self-estimated life expectancy | − No comparator prediction or measurement of survival − Only 3 of 105 patients had NYHA IV HF − Study methodology and tool not described − No index group without chronic disease was included |
| Wachterman et al20 2013 | ESRF | USA | Medium | Cross-sectional interviewer-administered questionnaire in two community-based haemodialysis units. | 62 patients receiving maintenance haemodialysis with 20% or greater predicted risk of dying in the next year. Mean age 68 (SD 10) 42% Male 52% Black |
| For 1 year survival prediction, patients were more optimistic in 64% of patient–nephrologist pairs, whereas nephrologists were more optimistic in only 10%. For 5 year survival prediction, patients were more optimistic in 69% patient–nephrologist pairs, whereas nephrologists were more optimistic in only 2% Only 6% of patients thought they had a <50% chance of being alive at 5 years, whereas actual survival at 23 months was only 56% | Patient expectations of 5-year mortality are higher than observed. Patients were significantly more optimistic about their survival than their nephrologists. Patients’ 1 year survival expectations were more consistent with actual survival than clinician estimates. Patients who expected to live longer were more likely to opt for life-extending treatments | − 88 of 150 eligible patients were excluded or refused to participate − No index group without chronic disease was included |
| Ambardekar et al21 2016 (abstract only) | HF | USA | Not rated | Cross-sectional report of self-estimated life expectancy. Methodology not reported. Subcomponent of multicentre prospective cohort study | 161 ambulatory patients with advanced HF from 10 American centres |
| 64% of patients identified by a physician to have ‘high-risk’ HF estimated a life expectancy of >2 years. 40% died, were transplanted or required a mechanical left-ventricular assist device over a mean follow-up of 13 months | Patients expectations of outcome were optimistic compared with physician-predicted or observed outcomes | + Multicentre prospective cohort − Abstract only at time of review − No index group without chronic disease was included |
| O'Donnell et al22 2015 (abstract only) | HF | USA | Not rated | Self-assessment of prognosis in single centre cohort of hospitalised patients with HF. Methodology incompletely reported | 23 participants Mean age 73 66% Male 77% White | Patient self-assessment of life expectancy | 70% of patients estimated a life expectancy of >5 years 43% of patients estimated a life expectancy of >10 years | Self-estimated life expectancy probably exceeds likely outcomes, but no comparator data were available. Patients who did not want to discuss prognosis all expected to live >10 years | − Very small numbers − Abstract only at time of review |
CKD, Chronic kidney disease; COPD, chronic obstructive pulmonary disease; ESRF, end-stage renal failure; HF, heart failure; LVEF, left ventricular ejection fraction.