The Effectiveness of Shared Care in Cancer Survivors—A Systematic Review

Population	Intervention	Context	Outcome
cancer (MeSH)	shared care (MeSH)	follow-up (MeSH)	All outcomes are included.
neoplasms (MeSH)	co-management	After care (MeSH)
cancer*	“sharing of care”	aftercare
neoplas*	“collaborative care”	follow up
malignan*	“care coordination”	followup*
carcinoma*	“coordinated care”	postsurgery
sarcoma	“referral and consultation”	post-surgery
oncolog*	“cooperative behavio?r”	postsurgical*
tumo?r*	“shared service*”	post-surgical*
adenocarcinoma*	“delivery of health care”	postoperat*
infiltrat*	“integrated care”	post-operat*
medullary	“shared model”	“continuity of patient care”
intraductal	“inter-organizational coordination”	“disease management”
		surveillance
		“disease progression”
		survivorship
		rehabilitation
		post treatment
		post-treatment
		posttreatment

Table 2

Overview of the included randomised controlled studies.

Author, year, country	Sample and setting	Intervention group	Control group	Length of follow-up, inclusion time	Cancer type	Outcomes	Conclusion
Bergholdt, et al. [31, 32], 2012, 2013, Denmark	955 newly diagnosed cancer patients (≧18 years) treated in a public regional hospital from May 2008 to February 2009 and 323 Danish general practices (GP) were enrolled to the trial.	n = 486, two nurses with oncological experience invited the patients to join an interview about rehabilitation needs, and patients were suggested to consult their GP when necessary.	n = 469, usual care	months, all patients were diagnosed as cancer within the previous 3 months	All cancers except carcinoma in situ and non- melanoma skin cancers	No significant differences between two groups with EORTC QLQ-C30 survey at 6 months or 14 months, and POMS at 14 months. No significant differences between two groups with the number of GP proactive contact nor patients’ participation in rehabilitation activities.	The multimodal intervention could not benefit patients with health-related quality of life or psychological distress. It could not improve GP proactive contact nor patients’ participation in rehabilitation activities, neither.
Nielsen, et al. [33], 2003, Denmark	248 cancer patients (≧18 years) from the oncology department of a university hospital between August 1998 and December 1998.	n = 127, the intervention includes three elements: provide GP and patients with standardised knowledge package, build the communication channels between hospital and GP, and encourage patients to contact their GPs when they had healthcare problems.	n = 121, normal procedure	6 months, both primary cancer and recurrent cancer patients were included	All types of cancer except lymphoma	Patients’ attitudes toward the shared care and their GPs is more positive in the intervention group at 3 months in 2 of 4 items (p < 0.05), and patients contacted with their GPs more than the control group (p < 0.05). However, no significant differences were detected in quality of life or performance status between two groups. The subgroup analyses indicated that men and young patients (18–49 years) tent to rate the shared care more positively.	Shared care could enhance the involvement of primary care and patients, which influenced the patients’ attitude toward healthcare services positively, especially for men and young patients.
Holtedahl, et al. [34], 2005, Norway	91 adult cancer patients (≧18 years) in a university hospital were enrolled from October 1999 to September 2000.	n = 41, two 30 minutes visits were conducted by GP and physician separately. The topic is about patients’ experiences related to cancer. Besides, patients were encouraged to contact their GPs if they had any health-related problem.	n = 50, not mentioned about the care content.	6 months, both primary cancer and recurrent cancer patients who had finished cancer therapy were included	All types of cancer except pre- cancerous conditions such as in situ cervical cancer	There was no significant difference between two arms in quality of life, patients’ satisfaction, or the consultation times with GP at six months. The relatives’ satisfaction improved in the intervention group at six months (P = 0.018).	Some patients might benefit from the GP follow-up, but the improvements were not identified this study. Besides, the intervention did not generate closer contact between patients and their GPs.
Grunfeld, et al. [35], 2006, Canada	968 early-stage breast cancer patients from six cancer centres were enrolled from January 1997 to June 2001.	n = 483, the family physician received a one-page guideline from the cancer centre. The guideline included follow-up arrangement, referral instruction, and disease relevant information. The patients received follow-up care from their family physician.	n = 485, the patients received follow-up care from cancer centre	From less than 2 years to 5 years, patients who had been diagnosed with cancer from 9 to 15 months ago and had completed treatment 3 months ago	early stage breast cancer	The percentage of recurrence, death, and serious clinical event in shared care group were 11.2%, 6.0%, and 3.5%, respectively, compared to 13.2%, 6.2%, and 3.7%, respectively, in control group. The health-related quality of life that assessed by SF-36 and HADS between two arms did not show significant differences.	The shared care follow-up did not increase the risk of important recurrence-related serious clinical event or health-related quality of life for breast cancer patients.
Blaauwbroek, et al. [36], 2008, Netherlands (historical control design)	121 adult survivors (≧18 years) who used to be diagnosed as childhood cancer in a paediatric oncology department and did not join any follow-up study agreed to enter the study in 2004 and 2005.	n = 121, three visits were carried out during three years period. Visit 1 and visit 3 were conducted by an on-site family doctor at the medical centre, visit 2 was conducted by the local family doctor. Assessments, survey, and individualised follow-up suggestion were provided in these visits.	The data from another matched study in the Netherlands was used as control group data.	years, all patients had been treated in the hospital at least 5 years ago	All childhood cancers and Langerhans-cell histiocytosis, except central nervous system tumours	At visit 1, patients in intervention arm showed lower level of health-related quality of life with RAND-36 survey compared to control group data, but no significant differences were detected at visit 3. Patient satisfaction was assessed at visit 2, and 89 (88%) of the 101 patients were satisfied with shared care model.	Shared care involved paediatric oncologists and family doctors can be used in the long-term follow-up in childhood cancer patients.
Emery, et al. [37], 2017, Australia	prostate cancer patients with low risk or moderate risk features were recruited from two rural and four urban medical centres from November 2011 to July 2013.	n = 45, a GP visit was conducted to re-engage their relationship with patients. Besides, the treatment centre provided a SCP to the patients and their GPs, and a register and recall system was built to enhance GPs’ and patients’ compliance. After that, GP replaced two hospital routine follow-up visits at 6 months and 9 months.	n = 43, usual care (the patients visited the hospital specialists every three months)	months, all patients had completed curative intent treatments within the previous 8 weeks	Prostate cancer with low risk or moderate risk features, expect the metastatic disease	No significant differences between two groups with the HADS, the CaSUN, the EPIC, and the PSQ-18 results. But the intervention group had a preference of shared care than control group (P < 0.001), and shared care costs less than usual care.	Shared care model could not benefit prostate cancer patients with distress, unmet needs, cancer-specific quality of life, or satisfaction, but it is feasible as to provide a similar outcome with less money.
Mayer, et al., [38], 2016, United States	37 cancer patients (≧21 years) who had completed curative intent treatment in a cancer hospital were recruited.	n = 20, the research nurse in the hospital made a draft SCP, then the oncology nurse practitioner made a 40 minutes transition visit with the patient and revised the SCP. The final SCP was sent to both patients and their PCPs. Besides, the PCPs received relevant knowledge about follow-up. Afterwards, the PCPs carried out a semi-structured follow-up visit with patients within four weeks.	n = 17, the same with intervention group but without GP visit	6 weeks, all patients had completed curative intent treatments within the previous 4 to 6 weeks	All types of cancer except metastatic cancer	All patients reported having less contradictory information about care (P < 0.0001) and expected less follow-up from oncologists (P = 0.03) at 6 weeks. All PCPs felt more confident (P = 0.01) about survivorship care. However, the PCP visit has little effect on the results, except making a borderline difference with worry level (P = 0.05).	The SCP could improve patients’ and PCP’s confidence in survivorship information, but no significant benefit was identified about GP visit. This is a pilot study, a larger study should explore the changes in the future.

[i] GP = general practitioner. EORTC QLQ-C30 = the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30. POMS = Profile of Mood States Scale. RAND-36 = RAND 36-Item Health Survey. SCP: survivorship care plan. HADS = 14-item Hospital Anxiety and Depression Scale. CaSUN = Unmet needs were assessed with the Cancer Survivors’ Unmet Needs measure. EPIC = Expanded Prostate Cancer Index Composite. PSQ-18 = 18-item Short-form Patient Satisfaction Questionnaire. SF-36 = Medical Outcomes Study Short Form 36-Item General Health Survey.

Table 3

Overview of the other studies.

Author, year, country	Research design	Sample and setting	Procedures (Intervention and comparison)	Length of follow-up, onset time	Cancer type	Outcomes	Conclusion
Hanan, et al., [29], 2014, Ireland	Descriptive, single-centre, mixed methods	The community nurse took the responsibility of management of cancer patients instead of hospital nurses, and the visit was conducted at patients’ home.	The research included a six months skills training (from specialist cancer staff) for community nurses, specific referral form, and provided hospital support for community nurses in urgent situation by phone. The patients were told that the medical oncologists still took the responsibility of them. The patients were majorly managed by community nurses.	5 years, not mentioned	Not mentioned	Quantitative outcomes: Community nurse-led cancer care was compared with hospital activity data before the programme (baseline), in the middle of the programme (after the six months training), and five years after the baseline. The community nurses delivered safe cancer care, and the hospital attendances was decreased. The community nurses had more confidence in cancer management. Qualitative outcomes: The patients reported a shorter travel to health care, and improved quality of life. They had confidence in community nurses, and they indicated their sense of autonomy increased.	Shared care approach with training community nurses could benefit both patients and health care providers.
Blaauwbroek, et al. [39], 2012, Netherlands	Descriptive, single-centre, quantitative	80 childhood cancer survivors (≧18 years) who used to be diagnosed in a long-term follow-up clinic and did not join any follow-up study and their family doctors (n = 79) joined this study from September 2008.	The research team (hospital specialists) constructed a personalised SCP and sent the printed booklet to the patients. The plan was also accessible through a secure website to the survivors and their family doctors. Besides, the research team asked the survivors to make a half-hour meeting with their family doctors, they would remind the survivors again if they did not meet their family doctors six months later. Follow-up was done primarily by the family doctor.	one year, at least 5 years off-treatment	All types of cancer expect central nervous system tumours, survivors of all three levels of risk were included (low-risk, medium-risk, and high risk)	The family doctors were asked to finish an information form about the screening and the additional test results. Besides, an 18-item questionnaire and a 14-item questionnaire were used to assess survivors’ and family doctors’ views, respectively, about the website and the provided information, the late effects were also collected by the family doctors. Outcomes: 73 survivors and 72 family doctors completed the study. 97% of the survivors thought the follow-up was beneficial, but 11% of the survivors felt that the information was inadequate. 82% of the survivors believed in their family doctors’ ability in the follow-up. However, 77% of survivors felt worried about the reawakened memories about the disease. 60 family doctors (83%) finished all the recommended tests. 95% of the family doctors felt the follow-up was more beneficial and 83% felt their knowledge of late effects increased. Besides, 93% of the family doctors felt confident about management of follow-up if they received the SCP.	The shared care with web-based SCP is available in the long-term cancer follow-up of adults with childhood cancer, and most survivors and family doctors were satisfied with this model. Besides, both survivors and the family doctors thought their relevant knowledge improved though this process. The negative effect was the follow-up might reawaken survivors of bad memories about disease.
Berger, et al., [40], 2017, France	Descriptive, single-centre, quantitative	150 childhood cancer survivors (≧18 years) and their GPs (n = 106) involved in this study from December 2010 to June 2013. These survivors were from a region cancer registry and diagnosed of cancer before 15 years old.	The paediatric oncologist and the internist invited the survivors to join a consultation, in which the practitioners explained the medical history and potential late effects to the survivors. Additional tests were conducted based on the consultation. Besides, the medical doctor provided the consultation summaries and recommendations about follow-up (recommendation card) for the survivors and their GPs. The medical doctor would call the GPs if they did not respond to the study at first.	One year, diagnosed as primary cancer from January 1987 to December 1992.	All types of cancer expect leukaemia	120 survivors finished the nine-question satisfaction form, 107 of them (89%) were satisfied with the consultation. 86% of the survivors found the recommendation card was useful. More than 75% felt their lifestyle changed such as physical activity and diets. Type of cancer and treatment would influence the satisfaction with the follow-up (p < 0.05). The survivors who received chemotherapy felt more satisfaction with the shared care (p = 0.03). 106 GPs finished the five areas satisfaction survey, 63 of them (59%) reported they were not provided enough information about the patient’s treatment, 82 of them (77%) reported lack of late effect of chemotherapy. But most of them felt the recommendation card was useful (61%), over 80% of them appreciate the collaboration and availability of contact with hospital.	The long-term follow-up by collaboration of GP and hospital specialist was feasible and could benefit childhood cancer survivors and family physicians.
Lund, et al. [41], 2016, Denmark	Descriptive, multi-centre, quantitative	530 cancer patients from three hospitals were transferred to their GPs and invited to join the shared care follow up between September 2011 and March 2012, the follow-up lasts for three years.	The urological specialists developed follow-up recommendations for evaluating patient and GP compliance, including the schedule of follow-up and the instructions of referral. The patients were transferred to their GPs.	3 years, with or without relapse	Prostate cancer	426 (80.8%) patients were analysed, 390 patients (91.5%) were rated as “acceptable” or “good” compliance (the compliance was rated as four columns: unknown, poor, acceptable, and good); a total of 393 GP (92.3%) were rated as “good” or “acceptable” compliance (the compliance was rated as four columns: unknown, poor, acceptable, and good).	Shared care model is a safe approach with high rate of patient and GP compliance

[i] GP = general practitioner. SCP: survivorship care plan.

Table 4

Essence of shared care.

Author, year, country	Major follow-up care provider in shared care	Other follow-up care provider in shared care	Method of shared care involvement	Number of formal interaction between different care providers, communication tool between care providers	Number/frequency of interaction between patients and primary care in intervention group	Number/frequency of interaction between patients and primary care in control group	Difference between two groups
Hanan, et al. [29], 2014, Ireland	Community nurses	Oncology day-ward nurses, specialist cancer staff	The shared care included a six months skills training (from specialist cancer staff) for community nurses, specific referral form, and provided hospital support for community nurses in urgent situation by phone. The patients were told that the medical oncologists still took the responsibility of them.	More than once, skill training, phone, and resource book	Not mentioned	Not mentioned	Not mentioned
Bergholdt, et al. [31, 32], 2012, 2013, Denmark	General practitioner	Two hospital nurses with oncological experience	The hospital nurses suggested patients to consult their GPs when necessary. Equally, they also encouraged the GPs to be proactive to offer support to their patients, and send the GP an email, which include the patients’ information.	Once, Email and phone	Patients reported: 168 contacts (101 GP proactive contact, 61.1%)GP reported:379 contacts (232 GP proactive contact, 61.2%)	Patients reported: 156 contacts (81 GP proactive contact, 51.9%) GP reported: 373 contacts (206 GP proactive contact, 55.2%)	No significant differences in both patients reported and GP reported contacts.
Nielsen, et al. [33], 2003, Denmark	Both hospital specialists and GPs	/	Transferring knowledge and information from hospital to GP, building communication channels, encouraging patients to contact their GPs	Once, ordinary mail	Not mentioned	Not mentioned	More patients had contact with their GP in intervention group (p = 0.049 at 3 months, p = 0.046 at 6 months)
Holtedahl, et al. [34], 2005, Norway	Both GP and hospital physician	/	The GP was asked to initiate a consultation with patients. The patients were encouraged to contact their GPs if they have health related problem.	Once, unclear	The average consultation time was 1.26 per patients	The average consultation time was 1.04 per patients	No significant difference between two arms
Grunfeld, et al. [35], 2006, Canada	Family physician	Practitioner in cancer centre	The family physician received a one-page guideline from the cancer centre. The guideline included follow-up arrangement, referral instruction, and disease relevant information.	Once, unclear	Not mentioned	Not mentioned	Not mentioned
Blaauwbroek, et al. [36], 2008, Netherlands	Both paediatric oncologists and family doctors	/	The medical centre advised the patients to meet their family doctors. The patient information and the test results were shared by medical centre and primary care.	More than once, email or telephone	Not mentioned	Not mentioned	Not mentioned
Emery, et al. [37], 2017, Australia	Both hospital specialists and GPs	/	The treatment centre faxed a SCP to the GP. The register and recall system sent GP follow-up reminder letters.	More than once, fax, letter	Not mentioned	Not mentioned	Not mentioned
Mayer, et al. [38], 2016, United states	Both hospital nurses and GPs	/	Hospital nurses provided SCP to both patients and PCPs, they also conducted a transition visit with the survivors. The PCP carried out a semi-structured follow-up visit with survivors (talking points were developed by the hospital).	More than once, Electronic health system, mail, or email, websites	Not mentioned	Not mentioned	Not mentioned
Blaauwbroek, et al. [39], 2012, Netherlands	Family doctors	Long-term follow-up clinic	The research team (hospital specialists) constructed a personalised SCP website which was accessible to the family doctors. Besides, the research team asked the survivors to make a half-hour meeting with their family doctors, they reminded the survivor again if they did not meet their family doctors six months later.	More than once, secure website or letter	Not mentioned	Not mentioned	Not mentioned
Berger, et al. [40], 2017, France	GP	A paediatric oncologist and an internist	The medical doctor provided the consultation summaries and recommendations about follow-up for the survivors and their GPs. The medical doctor would call the GPs if they did not respond to the study at first.	Once, unclear	Not mentioned	Not mentioned	Not mentioned
Lund, et al. [41], 2016, Denmark	GP	Hospital outpatient urological specialists	The urological specialists developed follow-up recommendations for evaluating patient and GP compliance, including the schedule of follow-up and the instructions of referral.	Once, unclear	Not mentioned	Not mentioned	Not mentioned

[i] GP = general practitioner. SCP: survivorship care plan.

Table 5

Critical review of the RCTs with Critical Appraisal Skills Programme (CASP).

Section/Question	Score
Section/Question	Bergholdt, et al., 2012 [31]	Bergholdt, et al., 2013 [32]	Nielsen, et al., 2003 [33]	Holtedahl, et al., 2005 [34]	Grunfeld, et al., 2006 [35]	Blaauwbroek, et al., 2008 [36]	Emery, et al., 2017 [37]	Mayer, et al., 2016 [38]
(A) Are the results of the trial valid?
1. Did the trial address a clearly focused issue?	2	2	2	1	2	2	2	1
2. Was the assignment of patients to treatments randomised?	2	2	2	2	2	2	2	2
3. Were patients, health workers and study personnel blinded?	0	0	0	0	0	0	0	1
4. Were the groups similar at the start of the trial?	2	2	1	1	2	1	2	2
5. Aside from the experimental intervention, were the groups treated equally?	1	1	1	1	2	0	1	2
6. Were all of the patients who entered the trial properly accounted for at its conclusion?	2	2	2	2	2	2	2	2
(B) What are the results?
7. How large was the treatment effect?	2	2	2	1	2	2	2	1
8. How precise was the estimate of the treatment effect?	2	2	2	2	1	2	2	1
(C) Will the results help locally?
9. Can the results be applied in your context? (or to the local population?)	1	1	2	1	0	1	1	1
10. Were all clinically important outcomes considered?	1	2	2	1	2	2	1	1
11. Are the benefits worth the harms and costs?	2	2	2	2	2	2	2	2
Total score (maximum 22)	17	18	18	14	17	16	17	16

[i] “0” represents many limitations, “1” represents some limitation, “2” represents excellent.

Table 6

Critical review of the quantitative studies with Health Care Practice R&D Unit (HCPRDU).

Question	Score
Question	Blaauwbroek, et al., 2012 [39]	Berger, et al., 2017 [40]	Lund, et al., 2016 [41]
(1) Study overview	2	2	1
(2) Study, setting, sample and ethics	2	2	1
(3) Ethics	2	0	1
(4) Group comparability and outcome measurement	1	1	1
(5) Policy and practice implications	1	1	2
(6) Other comments	2	1	2
Total score (maximum 12)	10	7	8