Effect of pediatric palliative care programs on health care resource utilization and costs among children with life-threatening conditions: a systematic review of comparative studies

Tania Conte; Craig Mitton; Logan M. Trenaman; Negar Chavoshi; Harold Siden

doi:10.9778/cmajo.20140044

Article Figures & Tables

Figures

Tables

Download figure
Open in new tab
Download powerpoint
Figure 1:
Flow diagram of study selection.

Tables

Figures

View popup

Table 1: Characteristics of 11 studies comparing pediatric palliative care programs and usual care identified through the systematic review

Study	Participants	Study design	Observation period	n	Intervention group	Comparison group	Outcome measures
Postier et al., 2014¹⁸	Age 1–21 yr Enrolled in home PPCP/hospice program	RChBA*	Before: 1 yr After: 1 yr	425	After PPCP	Before PPCP	Change in no. of hospital admissions, LOS and total billed charges for hospital/ED stay
Fraser et al., 2013¹¹	Age ≤ 19 yr Death from cancer	RCS	Referral to death	497	Hospice group (n = 132)	Control group (n = 311)	Total no. of hospital admissions, no. of planned hospital admissions, no. of emergency admissions
Keele et al., 2013¹²	Age < 18 yr Death in hospital from any cause > 5 d after admission	RCS	Last admission before death	24 342	PC group (n = 919)	No PC (n = 23 423)	Age, sex, LOS, major group category diagnostic, medications, procedures in last admission
Arland et al., 2013¹⁰	Age 1 mo–19 yr Death from brain tumour	RChBA†	Before: 5 yr After: 10 yr	114	After group (n = 92)	Before group (n = 22)	Symptoms, hospital admissions (no., LOS), location of death
Smith et al., 2013¹⁹	Children (age not defined) Hospital costs after discharge were in top tenth percentile	RCS + RChBA*	RCS: up to 2 yr RChBA: undisclosed	1 001	PPCP group (n = 81)	Control group (n = 920)	Cost, demographics, use of technology
Gans et al., 2012⁵	Age ≤ 20 yr Life-threatening condition Enrolled in PPCP	RChBA*	Before: 12 mo‡ After: 18 mo	123	After PPCP	Before PPCP	LOS, medical expenditures, family’s quality of life and satisfaction
Pascuet et al., 2010¹³	Children (age not defined) Used respite at pediatric hospice	RChBA*	Before: 1 yr After: 1 yr	66	After respite	Before respite	LOS, ED and outpatient visits, overall cost in hospital/hospice admission
Dussel et al., 2009¹⁴	Children (age not defined) Death from cancer	XS survey + RCS	Last month of life	140	Location of death planned (n = 88)	Location of death not planned (n = 52)	End-of-life planning, end-of-life support from physicians, use of home care, hospital resource utilization, place of death
Knapp et al., 2009¹⁵	Age 1–21 yr Death from any cause	RCS	Last year of life	1 527	Hospice (n = 85)	No hospice (n = 848)	Hospice use, hospice expenditures, other expenditures
Ward-Smith et al., 2008¹⁷	Children (no age bracket defined) Enrolled in PPCP	RCS§	6 mo before death	18	PPCP group (n = 9)	Non-PPCP group (n = 9)	Total hospital costs, LOS, differences in types of procedures
Belasco et al., 2000¹⁶	Children (no age bracket defined) Referred to home PPCP	Case series	1 d	3	Home care	Hospital care	Type of interventions delivered, place of death, comparison of charges of care

Note: ED = emergency department, LOS = length of stay, PC = palliative care, PPCP = pediatric palliative care program, RChBA = retrospective cohort before–after study with no control, RCS = retrospective cohort study, XS = cross-sectional study. *The same patients were followed before and after the intervention with no controls. †Different cohorts of patients were followed before and after the intervention (historical cohort comparison). ‡Length of “before” follow-up period unclear. §Described by the authors as a case–control study but technically it was a cohort comparison.

View popup

Table 2: Summary of quality assessment

Study*	Newcastle–Ottawa Scale quality assessment score
Study*	Selection (max 4 stars)	Comparability (max 2 stars)	Outcome (max 3 stars)
Postier et al.¹⁸	★ ★	★	★ ★ ★
Fraser et al.¹¹	★ ★ ★	★	★ ★
Knapp et al.¹⁵	★ ★ ★	★	★ ★
Keele et al.¹²	★ ★ ★	★	★
Arland et al.¹⁰	★ ★ ★		★ ★
Dussel et al.¹⁴	★ ★ ★		★
Ward-Smith et al.¹⁷	★		★ ★
Gans et al.⁵	★	★	★
Pascuet et al.¹³	★ ★		★

*The case series of 3 patients¹⁶ and the conference abstract¹⁹ were not included in the quality assessment.

View popup

Table 4: Effect of pediatric palliative care programs versus usual care on length of stay

Study*	Summary	Outcome
Postier et al.¹⁸	Decrease	Mean total LOS ± SD: 34.09 ± 59.7 d before v. 19.37 ± 34.0 d after intervention; p < 0.001 Interaction: level of exposure to program, noncancer and study period (p < 0.001)
Arland et al.¹⁰	Decrease	Mean LOS per hospital admission: 3.03 after v. 4.05 d before intervention; 25% decrease (no test applied) Mean LOS per patient overall (not only among those admitted to hospital): 1.25 d after v. 3.68 d before intervention; 66% decrease (no test applied)
Dussel et al.¹⁴	No difference/ controversial	Median (IQR) LOS in last mo of life: 17 (4–27) d in intervention group v. 21 (6–28) d in control group; p = 0.494
Ward-Smith et al.¹⁷	No difference/ Controversial	Mean LOS (range): 4 (5 to 17) d in intervention group v. 4 (5 to 18) d in control group
Gans et al.⁵	Decrease	Mean LOS per patient per mo: 4.0 before v. 2.8 after intervention; 32% decrease (no test applied)
Pascuet et al.¹³	Increase	Median no. of days in hospital per mo (range) after intervention: –2.9 (95% CI –4.5 to –1.3); p = 0.001 Median no. of days in hospice per mo (range) after intervention: 2.4 (0.08–26.5) Variation in median no. of total inpatient days per mo after intervention: 0.9 (p = 0.013)

Note: CI = confidence interval, IQR = interquartile range, LOS = length of stay. *Studies listed according to outcome measurement and quality-assessment rank.

View popup

Table 3: Effect of pediatric palliative care programs versus usual care on admissions

Study*	Summary	Outcome
Arland et al.¹⁰	Decrease	54% with admissions before program (5 yr) v. 29% after program (10 yr); p < 0.05 46% fewer admissions
Dussel et al.¹⁴	No difference/ controversial	54% (47/87) with admissions in intervention group v. 98% (51/52) in control group; p < 0.001 No difference in emergency department visits: 6% (5/84) v. 16% (8/50); p = 0.057
Keele et al.¹²	Decrease	Relative risk of ICU admissions lower in intervention group: 0.29 (95% CI 0.26–0.32)
Smith A et al.¹⁹	No difference/ controversial	PICU admissions: 90% (73/93) of patients in intervention group v. 56% (522/920) in control group; p < 0.001 NICU admissions: 17% (14/93) in intervention group v. 28% (262/920) in control group; p = 0.04
Postier et al.¹⁸	No difference/ controversial	Mean no. of admissions ± SD: 3.09 ± 3.6 before v. 3.18 ± 4.3 after intervention; p = 0.538 Interaction: level of exposure to program, cancer or not, study period (p < 0.001), adjusting for other demographic and clinical characteristics
Fraser LK et al.¹¹	No difference/ controversial	Total hospital admissions (controlled for age, disease, sex, deprivation category): IRR 0.79 (CI 95% 0.59 to 1.05); p = 0.10 Planned hospital admissions: IRR 0.60 (CI 95% 0.43 to 0.85); p = 0.004 Emergency department visits: IRR 1.15 (CI 95% 0.84 to 1.58); p = 0.375
Pascuet E et al.¹³	No difference/ controversial	Median no. of emergency department visits per month after intervention: –0.03 (95% CI –0.09 to 0.02); p = 0.20 Median no. of outpatient visits per month after intervention –0.5 (95% CI –1.0 to –0.05): p = 0.029

Note: CI = confidence interval, ICU = intensive care unit, IRR = incidence rate ratio, NICU = neonatal intensive care unit, PICU = pediatric intensive care unit. *Studies listed according to outcome measurement and quality-assessment rank.

View popup

Table 5: Effect of pediatric palliative care programs versus usual care on costs

Study*	Summary	Outcome
Postier et al.¹⁸	Decrease	Mean charges ± SD (factor of $10 000): 20.97 ± 43.3 before v. 10.91 ± 21 after intervention; p < 0.001 Interaction: level of exposure to program, cancer/noncancer, study period (p < 0.001)
Knapp et al.¹⁵	Increase	Total mean expenditures (hospice v. nonhospice users): All causes of death: $83 719 v. $36 597 Perinatal: $112 723 v. $78 218 Chronic: $84 047 v. $66 584 [Note: The total cost is the sum of the mean expenditures per type (inpatient, outpatient, hospice and pharmacy). For more details visit the primary article.]
Keele et al.¹²	Decrease	Median (IQR) daily charges, adjusted for geography: $9348 ($6070–$15 318) in intervention group v. $11 806 ($8017 to $18 352) in control group; p < 0.001
Smith et al.¹⁹	Increase	Change in median daily cost (IQR) within intervention group: $3827 ($2.9K–3.9K) before v. $4013 ($3.1K–5.1K) after intervention; p = 0.06 Median daily cost (IQR) in 2010/11: $3.8K ($3.1K–$4.7K) in intervention group v. $3.4K ($2.7K–$4.5K) in control group; p = 0.001
Ward-Smith et al.¹⁷	No difference/ controversial	Mean hospital costs (range), excluding surgical charges: $78 780 ($33 283 to $130 970) in intervention group v. $81 780 ($28 970 to $135 432) in control group
Gans et al.⁵	Decrease	Mean total medical expenditures per patient per mo: $15 653 before v. $13 976 after intervention 11% decrease overall: nearly $1 million in 18 mo (35% decrease inpatient care costs [$3571]; 34% increase in outpatient care costs [$1398]; 35% increase in pharmaceutical costs [$495])
Pascuet et al.¹³	Decrease	Change in total inpatient costs per mo (hospital + hospice): –$4252 (95% CI –$953 to –$7551) after intervention; p = 0.012 Unit costs per d: $2007 hospital inpatient v. $500 hospice
Belasco et al.¹⁶	Decrease	Mean charges per d (hospital v. home): Acute myeloid leukemia (toddler): $4283 v. $17 Neuroblastoma (preschool child): $2300 v. $325 Multiple chronic diseases (teenager): $8258 v. $1308

Note: CI = confidence interval, IQR = interquartile range, SD = standard deviation. *Studies listed according to outcome measurement and quality-assessment rank.