Skip to main content

Table 4 Summary table of included quantitative studies

From: When to stop? Decision-making when children’s cancer treatment is no longer curative: a mixed-method systematic review

Reference

Objective

Study design

Participants

Setting

Results

Methods/Quality

Maurer SH et al. 2010. [24]

Compare the parental self-reported rationale about treatment decisions.

Cross-sectional study.

Parents (n = 62) of children (n = 58) whose disease had progressed to the terminal stage.

1 Hospital: USA

Parental reasons for:

Interview questions pilot tested.

─Do not resuscitate status or terminal care: concern with quality-of-life (74%) and patient wishes (67%).

Trained interviewers.

Private and separately semi-structured interviews within 72 hours of participating in non-curative treatment decisions.

A convenience sampling technique.

Most parents were women.

No sample size estimation.

─Phase I Research Controlled Trial: the need to continue cancer-directed treatment (71%).

Reasons for both groups of treatment:

Semantic content analysis.

Treatment decisions:

The children of these parents ranged in age from 0.6 to 21.6 years; median, 11.4 years).

─Do not resuscitate or terminal care (47%),

To make a decision that did right by their child.

─Phase I Research Controlled Trial (53%),

Parents were the only decision-makers.

To take into account the medical facts.

To preserve the child’s legacy.

Tomlinson D et al. 2011. [25]

To identify the factors influencing decision making about treatment options for end-of-life.

Cross-sectional study.

One parent per child (n = 77; response rate = 67%) of children at end-of-life.

1 Hospital:

Parental factors: hope, increased survival time, and child’s quality-of-life.

Interview questions pilot tested.

Canada

Trained interviewers.

Single face to face interviews. All items of the questionnaire were closed ended with categoric responses.

Healthcare professionals’ factors: financial considerations and parent opinion.

No sample size estimation.

No probabilistic sampling.

Median age of children was 8.6 years (range, 6.2-13.3 years).

Response rate of healthcare professionals = 100%

Treatment decisions:

─Palliative cytotoxic chemotherapy:

Most parents were women (78%).

Parents (n = 42, 54.5%),

Univariate logistic regression analysis

Healthcare professionals (n = 20, 15.6%).

Hypothetical scenario was presented to parents and to healthcare professionals.

─Supportive care alone:

Parents (n = 35), Healthcare professionals (n = 108).

Healthcare professionals (n = 128):

Primary oncologist (n = 25).

p < 0.0001

Nurses (n = 97)

Factors of options given to them:

Social workers (n = 6).

Child’s quality-of-life.

Most healthcare professionals were women (85%).

Survival time.

Probability of cure.

Wolfe J et al. 2000 [26]

To describe the primary goal of cancer-directed treatment during end-of-life period.

Cross-sectional study.

One parent per family (n = 103; response rate = 72%) of children who died of cancer (median 3 years; range 1.1-8.0, years after death).

2 Hospitals: USA

Parental goal:

Interview questions pilot tested.

─To extent life (n = 87, 84%)

Trained interviewers.

Oncologist goal:

─To lessen suffering (n = 18, 42%).

No sample size estimation.

All items of the questionnaire were closed ended with categoric responses.

No probabilistic sampling.

(k = 0.16; 95% CI -0.11 - 0.42)

Regression analysis

Children mean age 10.8 years; SD, 6.7 years old at death.

The majority of parental interviews were administered by telephone.

Most parents were women (86%).

Primary oncologist (n = 42; response rate was not declared). Most of them were men (69%).

Only 16% of children participated in the decision-making process.

De Graves S et al. 2002 [27]

To explore the shift from cure to palliation.

In-depth history audit of medical records.

Medical records of 18 (64%) children who died of cancer. They ranged in age from 2 to 17 years; median age at death 10 years old).

1 Hospital:

For many families the hope of cure continued until the child was close to death.

Pretesting of the audit form used is not reported.

Australia

No probabilistic sampling technique.

They continued to seek curative or life extending options.

No sample size estimation.

Parents participated in the decision-making.

Content analysis technique.

Bell CJ et al. 2010 [28]

To explore the experiences in adolescents dying from cancer, including end-of-life discussions.

Retrospective review of medical charts.

103 medical charts from adolescents who

1 Hospital: USA

Timing of end-of-life discussions occurred very

Pretesting of the instrument used is

Parents were directly involved in the end-of-life decision.

died of cancer.

 

close to death for a significant number of adolescents.

not reported.

No sample size estimation.

No probabilistic sampling.

Children mean age at death 14.4 years; SD, 2.9 years old.

Univariable analysis.

Hilden JM et al. 2001 [29]

─ To explore perceived barriers to the delivery of end-of-life care.

A mailed questionnaire survey

228 (55%) paediatric oncologists responded the survey. Gender distribution was not reported.

All members of American Society of Clinical Oncology in the USA, Canada and the UK.

Barriers:

Validated instrument

Multivariate analysis

─ Family’s unrealistic expectations for cure (n = 98, 43%).

─ Family denial of the illness as terminal (n = 63, 27.6%).

─ To describe the factors influencing decision-making.

─ Family conflicts (n = 24, 10.5%).

Factors influencing d-m:

─ Absence of effective therapy (n = 213, 93.4%)

─ Request by patient/parent(s) to stop treatment (n = 198, 87%)

Mack JW et al. 2005 [30]

To identify the determinants of high-quality care at the end-of-life for children, as perceived by parents and physicians.

Cross-sectional survey.

One parent per family (n = 144; response rate = 65%) of children who had died of cancer (a mean of 3.2 years after death).

2 Hospitals: USA

The parents’ principal determinant was doctor-patient communication.

Interview questions pilot tested.

Trained interviewers.

No sample size estimation.

All items of the questionnaire were closed ended with Likert scales.

Physicians’ care ratings depend on biomedical rather than relational aspects of care.

No probabilistic sampling.

Recall bias.

Selection bias.

Multivariable analysis.

Median age of children at death was 8.9 years (range, 0.3-25.3 years).

No association was found between parent and physician care ratings (p = .88).

The majority of interviews were administered by telephone.

Most parents were women (83%).

Child’s primary oncologist (n = 52; response rate = 100%), most of them were men (65%).

All participants were directly involved in the end-of-life decision.

Edwards KE et al. 2008 [31]

To explore how closely mothers’ and fathers’ understandings of prognosis and treatment goals during the child’s end-of-life period were aligned.

Cross-sectional survey.

Pairs of mothers and fathers (n = 38; response rate = 56%) were interviewed an average of 4 years after their child’s death.

2 Hospitals: USA

During end-of-life, the lessening of suffering was the main primary treatment goal reported. However, only 34% of couples agreed on this goal (k = 0.07; 95% CI, 0.20 to 0.44). During the last month of life, 42% of parents concurred regarding lessening suffering (k = .0.35; 95% CI, 0.05 to 0.65). Among discordant pairs, there was no parental gender preference for a particular goal.

Interview questions pilot tested.

Trained interviewers unclear.

The majority of interviews were administered by telephone.

No sample size estimation.

No probabilistic sampling.

Bivariate analysis.

Median age of children at death was 10.3 years (range, 0.9-24 years).

The items of the questionnaire were closed ended, yet some of them requested open-ended elaboration.

All participants were directly involved in the end-of-life decision.

Hechler T et al. 2008 [32]

To investigate the bereaved parents’ perspective on end-of-life decisions.

Cross-sectional study.

Parents of 48 (Response rate = 35%) children who died of cancer (range 3 - 5, years after death).

6 Hospitals: Germany

64% discussed end-of-life decisions with the healthcare team.

Interview questions pilot tested.

Semi-structured, single interviews.

Trained interviewers

Depending on whether parents had had a discussion on decisions with the team, their decision on resuscitation differed.

No sample size estimation.

The majority of parental interviews were face-to-face.

No probabilistic sampling technique.

Children mean age 8 years at death, SD, 4.9 years old.

Fisher exact test

Most parents were women (94%).

Treatment decisions:

All participants were directly involved in the end-of-life decision.

─ Cancer-directed treatment (n = 18)

─ Do not resuscitate (n = 24)

─ Terminal care (n = 6).

Mack JW et al. 2008 [33]

To assess parent’s experiences who continued cancer-directed treatment after they had recognized that the child had no realistic chance for cure.

Cross-sectional study.

One parent per family (n = 53) of children who had died of cancer (a mean of 3.2 years after death).

2 Hospitals: USA

The main goals to continue cancer-directed treatment were: cure (20%), life extension (22%) and to lessen suffering (20%).

Interview questions pilot tested.

All items of the questionnaire were closed ended with categoric responses or Likert scales.

Trained interviewers.

No sample size estimation.

Response rate = 64%

31 parents reported that their child suffered as a result of cancer-directed treatment.

No probabilistic sampling.

Median age of children at death was 8.9 years (range, 0.3-25.3 years).

Multivariable analysis

29 reported that their child had received little to no benefit.

The majority of interviews were administered by telephone.

Most parents were women (83%).

All participants were directly involved in the end-of-life decision.

Hinds P et al. 1997 [34]

To identify the factors influencing decision making about treatment options for end-of-life.

Cross-sectional study.

Parents (n = 37, response rate = 44.6%) of children who had died of cancer (6─24 months after death).

1 Hospital: USA

Parental factors:

Interview questions pilot tested.

Information and recommendations given by healthcare professionals.

Trained interviewers.

Semi structured interviews were conducted via telephone with parents and face-to face with healthcare professionals. It was then followed up with a questionnaire with a Likert response option.

No probabilistic sampling.

Oncologist factors:

No sample size estimation.

Content analysis and

Children mean age at death 13.4 years; SD, 5.10 years old.

Patient and family preferences.

Treatment choices between curative and non curative measures.

Logistic regressions.

Patient’s prognosis and comorbid conditions.

Gender distribution was not reported.

Information and opinions from colleagues.

Healthcare professionals

 

(16 oncologists, 3 nurses, and 2 chaplains).

  

All parents were directly involved in the EOL decision.