Introduction
Methods
Search
Study selection criteria
Risk of bias and methodological quality evaluation
Data synthesis and analysis
Results
Study characteristics
Number of studies | Portion of total (%) | |
---|---|---|
Setting | ||
- Outpatient clinic | 20 | 90 |
- In- and outpatient clinic | 1 | 5 |
- Hospice | 1 | 5 |
Study design | ||
- Randomized controlled trial | 19 | 86 |
- Cohort study | 3 | 14 |
Types of cancer | ||
- Various cancer types (including breast, colorectal, lung, gynaecologic, head and neck, etc.) | 16 | 72 |
- Lung | 4 | 18 |
- Lymphatic | 1 | 5 |
- Prostate | 1 | 5 |
Number of patients in included studies | ||
- 1–100 | 5 | 22 |
- 101–250 | 10 | 45 |
- 251–500 | 4 | 18 |
- > 500 | 3 | 14 |
Control group received | ||
- No PROM (care as usual) | 15 | 68 |
- PROM | 7 | 32 |
Number of studies | Portion of total (%) | |
---|---|---|
PROMs | ||
- European Organisation for Research and Treatment of Cancer, Quality of Life - Cancer 30 (EORTC QLQ-C30 ) | 6 | 17 |
- European Organisation for Research and Treatment of Cancer, Quality of Life - Lung 30 (EORTC QLQ-LC13 ) | 3 | 9 |
- European Organisation for Research and Treatment of Cancer, Quality of Life - Breast 23 (EORTC QLQ-BR23 ) | 1 | 3 |
- European Organisation for Research and Treatment of Cancer, Quality of Life - Colorectal 38 (EORTC QLQ-CR38 ) | 1 | 3 |
- General symptoms (on a numeric scale) | 7 | 20 |
- Hospital Anxiety and Depression Scale (HADS) | ||
- Symptom Tracking And Reporting (STAR) | 3 | 9 |
- Supportive Care Needs Survey (SCNS) | 1 | 3 |
- Functional Assessment of Cancer Therapy – General (FACT-G) | 1 | 3 |
- Functional Assessment of Cancer Therapy – Prostate (FACT-P) | 1 | 3 |
- MD Anderson Symptom Inventory (MDASI) | 1 | 3 |
- Common Toxicity Criteria Adverse Events (CTCAE) | 1 | 3 |
- Chemotherapy Symptom Assessment Scale (CSAS) | 1 | 3 |
- Palliative Performance Scale (PPS) | 1 | 3 |
- Memorial Symptom Assessment Scale (MSAS) | 1 | 3 |
- Hospice Quality of Life (HQLI) | 1 | 3 |
- Center for Epidemiological Studies-Depression (CES-D) | 1 | 3 |
- Spiritual Needs Inventory (SNI) | 1 | 3 |
- Short Portable Mental Status Questionnaire (SPMSQ) | 1 | 3 |
- Therapy Related Symptom Checklist (TRSC) | 1 | 3 |
Location of data collection | ||
- Home | 7 | 32 |
- Outpatient clinic | 12 | 54 |
- In- and outpatient clinic | 1 | 5 |
- Not reported | 2 | 9 |
Methods of data collection | ||
- Paper | 6 | 27 |
- Electronic | 11 | 50 |
- Paper and electronic | 1 | 5 |
- Telephone | 2 | 9 |
- Not reported | 2 | 9 |
Feedback received by: | ||
- Health care professional | 15 | 68 |
- Patient and health care professional | 6 | 27 |
- Not reported | 1 | 5 |
Education in interpretation for health care professional | ||
- Yes | 6 | 27 |
- No | 14 | 64 |
- Not reported | 2 | 9 |
Risk of bias
PROM as intervention, with or without feedback to patients or health care professionals, compared with not using a PROM
Author and year | Survival/mortality | Morbidity | Health-Related Quality of Life (HRQoL) | Patient satisfaction | Process indicators |
---|---|---|---|---|---|
Basch et al. 2016 [26] | Overall survival after 1 year was higher in IG than CG; 75% vs 69%, p = 0.05. Difference was more pronounced among computer-inexperienced participants; 74% vs 60%, p = 0.02). Quality-adjusted survival (in months) observed in one year was higher in IG than CG; mean 8.7 months vs 8.0 months, p = 0.004 | NR | More patients showed improved HRQoL in IG than CG (34% vs 18%), and fewer patients reported declined HRQoL scores (38% vs 54%) | NR | Patients in IG were less frequently admitted to the emergency room (34% vs 41%, p = 0.02) or hospitalized (45% vs 49%, p = 0.08) Patient in IG received longer active chemotherapy compared with CG; mean 8.2 months (0–49 months) vs mean 6.3 months (0–41 months), p = 0.002 |
Davis et al. 2013 [29] | NR | NR | HRQoL did not significantly differ between groups. Mean scores were IG 92.5 (SD 12.3), CG 94.8 (SD 11.3); p > 0.10. | 85% of patients (n = 60) endorsed that all patients would benefit from an automated monitoring system to routinely assess their symptoms/HRQoL | Doctor/patient communication did not change over time in both groups |
Detmar et al. 2002 [16] | NR | NR | HRQoL did not significantly differ between groups on any of the subscales. On two subscales, a greater percentage of patients in IG compared with CG showed improvement over time (defined as 0.5 SD unit or greater change). Mental health (43% vs 30%; p = 0.04) and role functioning (22% vs 11%; p = 0.05) | Patient satisfaction was high in both groups. The degree of received emotional support was higher in IG than CG; mean 4.3 (SD 0.72) and 4.0 (SD 0.89); p = 0.05 Almost all patients (97%) reported that the HRQoL profiles provided an accurate picture of their functioning and well-being. 79% believed it enhanced their physicians’ awareness of their health problems | Doctor/patient communication was higher in IG than CG. Mean scores were 4.5 (SD 2.3) and 3.7 (SD 1.9), respectively; p = 0.01 HRQoL topics were discussed more often in IG than CG No differences were seen in patient management actions. Mean number of actions undertaken were are 0.6 in IG and 0.5 in CG A higher percentage of patients in IG than CG received counselling from their physician on how to manage their health problems 23% vs 16% p = 0.05 |
Hilarius et al. 2008 [30] | NR | NR | HRQoL did not significantly differ between groups. Specific data not supplied by authors | Patient satisfaction was high in both groups. No statistically significant group differences were observed. Specific data not supplied by authors | The mean composite communication score regarding HRQoL was higher in IG than CG (p = 0.009). Mean scores were 4.8 (SD 3.3) and 3.8 (SD 2.3), respectively In IG, HRQoL topics were discussed more frequently than in CG No differences were seen in patient management activities (referral, medication prescription, test ordering, modification chemotherapy) between groups |
Hoekstra et al. 2006 [31] | NR | After 2 months, f/u all symptoms, except coughing, were less prevalent in IG than in CG (range prevalence − 2.1 to − 24.3%). Only coughing was more prevalent in IG (14.9%).Constipation, vomiting and sleeplessness showed relatively large differences (24%, 18% and 18%, respectively) in favour of IG, but only constipation and vomiting were statistically significant (no specific data were given by authors) Fatigue, lack of appetite, shortness of breath and nausea were rated less severe in IG, but this was not statistically significant Severity of pain, coughing, sleeplessness and diarrhoea were rated equally severe in both groups. Only constipation and vomiting were significantly experienced as more severe (P < 0.05) | NR | NR | NR |
Kearney et al. 2009 [32] | NR | More patients in CG reported fatigues than in IG. 81.3% vs 67.3% respectively; odds ratio = 2.29 (95% CI 1.04 to 5.05) p = 0.040 Hand-foot syndrome was less often present in the CG compared with IG. 12.2% vs 24.0%, respectively, odds ratio = 0.39 (95% CI 0.17 to 0.92) p = 0.031 No differences were seen in vomiting, nausea, diarrhoea and sore mouth/throat More severe hand-food syndrome and distress were seen in IG compared with CG: mean 0.46 (SD 0.64) vs 0.22 (SD 0.49); p = 0.033) and mean 0.30 (SD 0.45) vs 0.16 (SD 0.34); p = 0.028, respectively Other symptoms showed no significant differences in severity and distress between groups | NR | NR | NR |
Matsuda et al. 2019 [43] | NR | NR | HRQoL did not significantly differ between groups. An effect size of 7.39 (95% CI − 6.39 to 21.17; p = 0.285) was seen in favour of the IG over time | NR | NR |
McMillan et al. 2011 [44] | NR | In both groups, depression scores declined significantly over time (p = 0.023). Decline of depression scores was more present in IG than in CG (p = 0.027) No between-group differences were seen in distress scores and spiritual needs Specific data not supplied by authors. | In both groups, HRQoL improved significantly over time (p < 0.001), but no differences between IG and CG were seen Specific data not supplied by authors | NR | NR |
Mills et al. 2009 [33] | NR | NR | IG had a lower overall and lung-specific HRQoL than the CG. Overall scores declined 6.6 (SD 12.5) in IG and inclined 0.2 (SD 15.7) in CG; p = 0.10 Lung-specific scores declined 6.3 (SD 14.9) in IG and inclined 3.5 (SD 18.4) in CG; p = 0.05. | Both groups reported high levels of satisfaction with their care. CG reported slightly higher satisfaction, and no significant associations were identified | Little participants of IG discussed their results with the HCP (23%, n = 13). Patient in IG discussed fewer topics with HCP’s than CG, no statistically significant differences were seen |
Rosenbloom et al. 2007 [36] | NR | NR | No significant differences were observed in HRQoL across the three study groups (p > 0.05) Mean score of IG, ACG and CG were 115.8 (SD 22.9), 113.3 (SD 24.5) and 112.2 (SD 21.4), respectively | No significant differences were observed in general satisfaction and satisfaction with communication across the three study groups (p > 0.05) General satisfaction mean scores of IG, ACG and CG were 22.4 (SD 4.2), 23.1 (SD 4.2) and 24.4 (SD 4.1), respectively General satisfaction mean scores of IG, ACG and CG were 21.2 (SD 2.8), 21.1 (SD 3.0) and 20.8 (SD 3.2), respectively | Change in clinical treatment did not significantly differ between groups |
Taenzer et al. 2000 [38] | NR | NR | Four HRQoL-subscales (emotional, cognitive, social and global functioning) did not significantly differ between groups. CG scored better on two HRQoL subscales: physical functioning (p < 0.05) and role functioning (p < − 0.01) compared with the IG There was a high degree of variation in scores, particularly on symptom scales, indicating a wide range of QoL | Satisfaction did not significantly differ between groups (P > 0.05). Overall levels of patient satisfaction were high | In IG, HRQoL topics were more frequently discussed than in CG. Number of topics mean 6.4 (SD 4.1) vs 2.5 (SD 2.9); p < 0.01 A higher percentage of taken actions on identified HRQoL topics was seen in IG than CG (73% vs 68.5%) |
Takeuchi et al. 2011 [39] | NR | NR | NR | NR | In IG and ACG, more symptoms were discussed than in CG; p = 0.040 and p = 0.08, respectively Number of discussed symptoms was higher for all groups at the first consultation compared with the third consultation; p = 0.004 |
Velikova et al. 2004 [40] | NR | NR | HRQoL was higher in IG compared with CG and ACG; p = 0.006 and p = 0.80, respectively. HRQoL was higher in ACG compared with CG; p = 0.01 Specific data not supplied by authors | NR | In IG, more HRQoL topics were discussed compared with CG; mean number of topics discussed were 3.3 vs 2.7 Consultations did not prolong in IG; mean time per consultation 12.6 min in IG vs 12.8 min in CG |
Velikova et al. 2010 [41] | NR | NR | NR | 86% (n = 85) of patients in IG perceived that PROMs were useful to tell physicians how they were feeling compared with 29% (n = 34) in CG. Between 79 and 89% of all patients rated their quality of care as ‘very good’ or ‘excellent’ | Communication in the IG was rated better than in the CG; p = 0.03. No significant differences were seen in communication between IG and ACG; p = 0.16 |
Williams et al. 2013 [42] | NR | NR | HRQoL increased by 3.31 points in the IG (p = 0.12), whereby an increase of 3.0 points was seen as clinically significant No specific data was supplied for CG by authors | NR | In IG, more symptoms were documented and managed than in CG; mean number of symptoms: 3.76 (p < 0.001) The number of symptoms documented and managed increased by 0.76 for each cancer stage greater than stage I (p < 0.03) |
Patient outcomes—survival/mortality
Patient outcomes—morbidity and symptoms
Patient outcomes—HRQoL
Patient experiences—patient satisfaction
Process indicators
PROM as intervention with feedback to patients or health care professionals, compared with a control group in which PROMs were used without giving feedback to patients or health care professionals about the results
Author and year | Survival/mortality | Morbidity | Health-Related Quality of Life (HRQoL) | Patient satisfaction | Process indicators |
---|---|---|---|---|---|
Berry et al. 2011 [27] | NR | NR | NR | NR | If symptoms or HRQoL issues reached the alert-threshold in the IG, there was a 29% increase in the odd that these symptoms or HRQoL were discussed in consultation; odds-ratio 1.287( 95% CI 1.047 to 1.583) Length of clinic visits did not differ in length between groups. IG mean 30.3 (SD 17.9) min vs CG 31.7 (SD 18.8) min |
Boyes et al. 2006 [17] | NR | Mean anxiety scores decreased in IG (6.83 at baseline to 4.80 at final f/u) more compared with CG (6.13 at baseline to 5.17 at final f/u); p = 0.09 Mean depression scores did not significantly differ between groups (p = 0.20). Mean depression scores in IG decreased from 4.98 to 4.20 (baseline to final f/u) and increased in CG from 3.84 to 3.91 (baseline to final f/u). No difference between IG and CG in moderate or high psychological needs (p = 0.82) | NR | NR | 34 of 36 patients rated the PROM as easy to complete. 30 of 36 patients thought that using a PROMS was a good way for doctors to get information about patients’ well-being 3 of 20 patient in IG reported that their physician discussed the feedback report with them (n = 3). Two of four HCPs reported they discussed the feedback report with their patients. Patients in IG were less likely to report a bothersome symptom at a third visit when they already reported it at the second visit, compared with CG; OR = 2.8, p = 0.04 |
Cleeland et al. 2011 [28] | NR | A significant reduction of symptom threshold events was seen in both groups. The reduction rate was 19% in IG and 8% in CG. Rate ratio difference was 0.88 (95% CI 0.78 to 0.98) indicating IG approximately had 12% less symptom threshold events | NR | Patient in the IG were more satisfied with the intervention than patients in CG; mean score: 9.4 vs 8.4 respectively, p < 0.03. Patients in the IG rated the system more likely as easy to use; mean score 9.7 in IG vs 8.8 in CG, p < 0.01 | NR |
Mooney et al. 2014 [34] | NR | Symptom severity and distress scores did not significantly differ between groups (mean difference = 0.06; p = 0.58). | NR | 79.0% of the patients in IG were quite or very confident that the automated system notified their physician of their symptoms 25.0% of the patients in IG agreed that the system helped their physician to decrease their symptoms | Unscheduled contacts did not significantly differ between groups (p = 0.73) Frequency of patient-initiated and physician-initiated contacts was similar (p = 0.14) Patients in CG talked somewhat more often about their symptoms (n = 79, 73.0%) at patient-initiated contacts than patients in IG (n = 64, 62.0%) There were more provider-initiated contacts that resulted in an office visit in the IG (n = 18, 17.5%) than in the CG (n = 10, 9.3%). In the provider-initiated contacts in IG symptoms were discussed more often (n = 14, 70.0%) than in the CG (n = 4, 33.0%), p = 0.10 |
Nicklasson et al. 2014 [35] | NR | NR | NR | NR | Emotional functioning was more discussed by doctors and patients in the IG than in the CG; mean 3.9 statements vs 2.4 statements; p = 0.015. Discussion of physical/role, social or cognitive functioning did not significantly differ between groups The sum of function-related statements by doctors and patients was higher in the IG compared with CG; mean 9.2 statements vs 6.9 statements; p = 0.0096 All symptoms (pain, dyspnea, fatigue, anorexia and other symptoms) were somewhat more discussed by doctors and patients in IG compared with CG (25.2 statements vs 24.5 statements), yet not significant, p = 0.36 Length of consultation was similar between groups. IG median 20 min vs CG median 22 min; range 8–60 min, p = 0.77) The number of diagnostic and therapeutic interventions per patient was statistically significant higher for emotional functioning (0.43 interventions vs 0.15 interventions; p = 0.0036), social functioning (1.17 interventions vs 0.74 interventions; p = 0.013) and dyspnea (1.08 interventions vs 0.53 interventions; p = 0.017), in the IG compared with the CG |
Ruland et al. 2010 [45] | NR | Symptom distress declined over time in 10 of 19 (58%) symptoms in the IG (pain, eating/drinking, bowel/bladder, energy, sleep/rest, concentration/memory, activities of daily living/self-care and worries/concerns). Symptom distress declined in 2 of 19 symptoms in the CG (pain and worries/concerns). Discomfort, eating/drinking, sleep/rest and sexuality statistically differed between groups in favour of the IG Specific data not supplied by authors | NR | NR | 17 of 19 symptoms showed a downward trend in patient needs for symptom management in the IG (p < 0.05) 14 symptoms in the CG showed an upward trend (6 of 19 were statistically significant p < 0.05) indicating that patients had greater needs for support to manage their problems over time Specific data not supplied by authors. |
Strasser et al. 2016 [37] | NR | Symptom distress score between first and last visit was statistically lower in IG compared with CG. Mean difference between IG and CG: 5.70 (95% CI 1.96 to 9.43); p = 0.003 | HRQoL was higher in IG than in CG. Mean difference between IG and CG 6.84 (95% CI − 1.65 to 15.33); p = 0.1 | NR | A trend favouring IG (p = 0.06) was seen in symptom management performance. 71 (52%) patients in IG vs 40 (38%) patients in CG had symptom management interventions in visits where their symptom load was above a pre-set threshold. Specific data not supplied by authors |
Patient outcomes—morbidity and symptoms
Patient outcomes—HRQoL
Patient experiences—patient satisfaction
Process indicators
Discussion
Conclusion
Acknowledgement
Compliance with ethical standards
Conflict of interest
-
Caitlin Graupner, Merel L. Kimman and Stéphanie Breukink received research grants from the Stichting Kwaliteitsgelden Medisch Specialisten.
-
All other authors declare that they have no conflict of interest.
-
Authors have full control of all primary data and allow the journal to review the data.