Review Question

Should integrated diagnostic reporting (via Specialist Integrated Haematological Malignancy Diagnostic Services [SIHMDS]) replace local reporting in the diagnosis of haematological malignancies?

What are the effective ways of delivering integrated diagnostic reports (for example, co-located or networked) in the diagnosis of haematological malignancies?

Background

The main driver for this recommendation in the improving outcomes guidance and subsequent 2012 revision (agreed by the National Cancer Action Team and the RCPath) was evidence of a significant misdiagnosis rate for haematological malignancies (5-15%) sometimes with major clinical consequences (Clarke et al., 2004; LaCasce et al., 2008; Lester et al., 2003; Proctor et al., 2011). This type of error can be difficult to detect after a patient has been treated and therefore a premium must be placed on being able to demonstrate that a diagnosis is correct and supported by strong evidence across several independent investigative modalities. This approach is intrinsic to the way that disease entities are defined in the World Health Organisation (WHO) classification and is common to all haematological malignancies.

The availability of the necessary investigations varies across the country. To be effective this multi-modality approach to diagnostic quality assurance requires a systematic approach to the investigation of specimens and a clear process to interpret and integrate the results obtained (via integrated diagnostic reporting), most crucially to identify inconsistencies between the results obtained by different investigative techniques. This is most effectively delivered within an integrated diagnostic service able to provide the full range of diagnostic techniques required and to provide a report to the end users that integrates these results into a single diagnostic assessment; this was the rationale behind the current guidance (Ireland et al, 2011). A very important subsidiary consideration is that diagnostic techniques are rapidly evolving and these developing techniques need to be reflected in laboratory organization. The efficient delivery of evolving modern diagnostic approaches, such as molecular genetics and flow cytometry, potentially across a range of specialities with the required quality and economy of scale needs to be balanced against the requirements of specialised integrated reporting, which, on a practical level, are easiest to achieve within a fully integrated laboratory or other closely located laboratory configuration. This is because the diagnosis, classification and prognostic assessment of these conditions requires integration of multiple diagnostic techniques and high levels of ascertainment and data quality can only realistically be achieved in an infrastructure which facilitates routine, direct interaction between component laboratory professionals.

High quality data on diagnosis, treatment and outcome data on cancer patients is a key objective of the NHS. Data quality in haematology has long been a major problem with widely differing levels of ascertainment between regions and the ability to report data in only the broadest categories of limited clinical utility. A greater implementation and standardisation of SIHMDS reporting should improve the quality of data in haemato-oncology and contribute to NHS goals. In addition, the integrated delivery of modern diagnostics in haemato-oncology is a highly active area of research and development that the NHS is uniquely placed to make an internationally competitive contribution.

However, there are a number of other important considerations for example, the availability of suitably trained staff (pathologists, clinical and biomedical scientists) is limited and constrains the number of centres able to offer this service. To ensure rapid diagnosis and to conserve diagnostic material (which in the case of needle core biopsies, may be sparse) it is important that specimens from patients suspected of having a haematological malignancy are referred directly to the specialist laboratory. This raises two problems, which have proved a significant obstacle to implementing this guidance. It is not always possible to identify specimens that require referral from the patient's clinical features alone and triage by local pathologist and haematologists is important. Concern is also expressed frequently that this means that local pathology staff will become deskilled and more broadly that referral of specialist work of this type undermines the viability and job satisfaction of local hospital laboratories. Although previous consensus recommendations have been made for minimum catchment populations for the delivery of SIHMDS (NCAT 2012), there is no evidence to support such thresholds. Delivery of SIHMDS may be influenced by regional configurations of clinical haematology and oncology services, including MDTs and academic networks, along with broader geographical considerations such as regional infrastructure and transport flows. Although Cancer Networks are no longer in operation, their effect may persist in NHS cancer services in regional working relationships and service delivery.

In recent multicentre UK studies, early mortality following AML induction chemotherapy has been reported as up to 6% and 9% at 30 days and 10% and 15% at 60 days in younger and older patients respectively (Burnett et al, 2015; Burnett et al, 2012).

Reported induction mortality is also substantial in ALL; 4% in patients <55 and 18% in patients over 55 years (Sive et al, 2012). Early mortality in ALL is not improved with the introduction of modern drugs, such as tyrosine kinase inhibitors in Philadelphia positive disease (Fielding AK et al, 2014). Recent data confirm a 2.2% induction death rate in 16-25 year olds treated on paediatric protocols. In 25 – 60 year olds treated on the current NCRI UKALL 14 type schedule, the induction death rate in UKALL 14 currently is 8.5% (personal communication, Dr Clare Rowntree).

Question in PICO format

Searching and Screening

Search Results

Total References retrieved (after initial sift and de-duplication): 270

Medline search strategy (This search strategy is adapted to each database)

1. exp Hematologic Neoplasms/
2. ((haematolog* or hematolog* or blood or red cell* or white cell* or lymph* or marrow or platelet*) adj1 (cancer* or neoplas* or oncolog* or malignan* or tumo?r* or carcinoma* or adenocarcinoma* or sarcoma*)).tw.
3. exp Lymphoma/
4. lymphoma*.tw.
5. (lymph* adj1 (cancer* or neopla* or oncolog* or malignan* or tumo?r*)).tw.
6. hodgkin*.tw.
7. lymphogranulomato*.tw.
8. exp Lymphoma, Non-Hodgkin/
9. (nonhodgkin* or non-hodgkin*).tw.
10. lymphosarcom*.tw.
11. reticulosarcom*.tw.
12. Burkitt Lymphoma/
13. (burkitt* adj (lymphom* or tumo?r* or cancer* or neoplas* or malign*)).tw.
14. brill-symmer*.tw.
15. Sezary Syndrome/
16. sezary.tw.
17. exp Leukemia/
18. (leuk?em* or AML or CLL or CML).tw.
19. exp Neoplasms, Plasma Cell/
20. myelom*.tw.
21. (myelo* adj (cancer* or neopla* or oncolog* or malignan* or tumo?r*)).tw.
22. kahler*.tw.
23. Plasmacytoma/
24. (plasm?cytom* or plasm?zytom*).tw.
25. (plasma cell* adj3 (cancer* or neoplas* or oncolog* or malignan* or tumo?r* or carcinoma* or adenocarcinoma*)).tw.
26. Waldenstrom Macroglobulinemia/
27. waldenstrom.tw.
28. exp Bone Marrow Diseases/
29. exp Anemia, Aplastic/
30. (aplast* adj an?em*).tw.
31. exp Myelodysplastic-Myeloproliferative Diseases/
32. exp Myeloproliferative Disorders/
33. exp Myelodysplastic Syndromes/
34. exp Thrombocytopenia/
35. (thrombocytop?eni* or thrombocyth?emi* or poly-cyth?emi* or polycyth?emi* or myelofibros or myelodysplas* or myeloproliferat* or dysmyelopoietic or haematopoetic or hematopoetic).tw.
36. exp Anemia, Refractory/
37. (refractory adj an?em*).tw.
38. (refractory adj cytop?en*).tw.
39. Monoclonal Gammopathy of Undetermined Significance/
40. (monoclonal adj gammopath*).tw.
41. (monoclonal adj immunoglobulin?emia).tw.
42. MGUS.tw.
43. ((oncohaematolog* or oncohematolog*) adj2 (disorder* or disease* or syndrome*)).tw.
44. or/1-42
45. limit 44 to yr=“2000 - 2015”
46. Clinical Laboratory Services/
47. Clinical Laboratory Information Systems/
48. Diagnostic Services/
49. (laborator* adj2 (service* or report*)).tw.
50. (laborator* adj1 (integrat* or central* or co-locat* or local* or region* or district* or communit* or hospital* or network* or specialis*)).tw.
51. (diagnos* adj2 (service* or report*)).tw.
52. (diagnos* adj1 (integrat* or central* or local* or region* or district* or communit* or hospital* or network*)).tw.
53. Pathology Department, Hospital/
54. Laboratories, Hospital/
55. Diagnostic Errors/
56. (diagnos* adj discrepanc*).tw.
57. (expert review* or expert patholog* review*).tw.
58. second review.tw.
59. central* review.tw.
60. ((haematopatholog* or hematopatholog* haematolog* or hematolog* or patholog* or histopatholog* or cytopatholog*) adj2 (service* or report*)).tw.
61. ((haematopatholog* or hematopatholog* haematolog* or hematolog*or patholog* or histopatholog* or cytopatholog*) adj1 (integrat* or central* or co-locat* or local* or region* or communit* or hospital* or network* or specialis*)).tw.
62. inter-laborator*.tw.
63. SIHMDS.tw.
64. exp laboratories/
65. Hospital Information Systems/
66. or/46-65
67. 45 and 66

Screening Results

Study Quality

A short checklist of relevant questions was developed to assess the quality of the included studies and from this it was judged that the included evidence was of low quality overall as all identified studies were retrospective case series studies and none of the included studies directly compared integrated diagnostic services with other forms of diagnostic services.

All studies included relevant populations with either general haematology patients or specific haematology subtypes such as lymphoma patients included in the individual studies.

Identified studies broadly compared the rates of discordance in diagnosis of haematological malignancies between initial diagnosis and review diagnosis by expert pathologists, sometimes based in a specialist laboratory, though it was unclear in the individual studies whether the expert pathologists were blinded to the initial diagnosis therefore there is a high risk of bias based on the potential lack of blinding.

The outcomes reported in each of the studies were not specifically those listed in the PICO table, however the outcomes reported (e.g. diagnostic discordance, change in management, survival) were considered to be of some use in informing discussions.

Overall, the quality of the evidence for this topic was considered to be low quality for all outcomes.

Evidence Statements

Low quality evidence from a total of nine retrospective studies of either haematology or lymphoma populations, two of which were UK based (Bowen et al, 2014; Chang et al, 2014; Herrera et al, 2014; LaCasce et al, 2005; Lester et al, 2003; Proctor et al, 2011; Siebert et al, 2001, Stevens et al, 2012, and van de Schans et al, 2013). The discordance rates between initial haematological pathological diagnoses and expert review ranged from 6%-60%. Revision of one type of lymphoma to another type was the most common source of discordance ,ranging from 6.5%-23% (2 studies; Bowen et al 2014; Chang et al, 2014).

Low quality evidence for major discrepancies, leading to a change in treatment or management was recorded in four retrospective studies (Chang et al, 2014; Lester et al; 2003; Matasar et al, 2012 and Stevens et al, 2012) with rate of discordance between an initial diagnosis and review diagnosis ranging from 17.8% to 55%.

Low quality evidence from one retrospective study (Engel-Nitz et al, 2014) which compared diagnostic outcomes between specialist haematology laboratories and other commercial laboratories and reported that patients in the specialist laboratory cohort were more likely to undergo more complex diagnostic testing with 26% of patients undergoing molecular diagnostics compared with 9.3% in community based hospital laboratories. Patients in the specialist laboratory cohort were 23% more likely to reach a final diagnosis within a 30 day testing period when compared with community based hospital laboratories.

Low quality evidence from one retrospective study compared a national registry of haematological malignancies with a hospital discharge registry to investigate the data quality and the impact of misclassification on survival in haematology patients (Norgaard et al, 2005). It reported the overall data completeness was 91.5% [95% CI, 89.6%-93.1%] and that the survival of patients registered in the hospital discharge registry was about 20% lower and about 10% lower for patients registered in the national registry when compared with patients registered in both.

Low quality evidence from a single retrospective study evaluating the value of expert pathology review (van de Schans et al, 2013) reported no statistically significant difference in 5-year survival between patients with a concordant diagnosis compared to those with a discordant diagnosis (48% [95% CI, 42%-53%] versus 53% [95% CI, 39%-67%]).

Low quality evidence from a retrospective study including 25 cases of Burkitt Lymphoma reviewed by 10 pathologists (Rane et al) reported a poor rate of concordance between the pathologists for independent diagnosis (κ0.168, SE±0.018) and a direct correlation between level of experience and diagnosis.Expert lymphoma pathologists showed marginally higher concordance rates and general pathologists the lowest (κ0.373 versus κ0.138). For consensus diagnosis the level of agreement between pathologists for revised diagnosis was very high (κ0.835, SE±0.021) and revision of diagnosis was highest among general pathologists. The concordance of independent diagnosis and consensus diagnosis was low (κ=0.259, SE±0.039; median=0.207; range=0.131-0.667) and increased with increasing experience of diagnosing lymphoma.

Low quality evidence from a retrospective study including 25 cases of Burkitt Lymphoma reviewed by 10 pathologists (Rane et al) reported that expert lymphoma pathologists were significantly more likely to make a correct diagnosis compared with both pathologists with experience (OR=3.14; p=0.012) and general pathologists (OR=5.3; p=0.00032).

Low quality evidence from two retrospective studies (Matasar et al 2012 and Strobbe et al, 2014) showed that the rates of discordance between initial and review diagnoses were found to have dropped between 2001 and 2005, but with no statistically significant difference. Matasar et al, 2012 reported a drop in major revision rates for haematological malignancies from 17.8% to 16.4% (p=0.6) as familiarity with the WHO classification system increased and Strobbe et al, 2014 reported a drop in discordance rate of lymphoma diagnoses from 14% to 9% (p=0.06) following the setting up of an expert lymphoma review panel.

Low quality evidence from two retrospective studies (Irving et al, 2009 and Norbert-Dworzak et al, 2008) reported that interlaboratory agreement was high for the use of a standardised protocol for flow cytometry (correlation coefficient ranged from 0.97-0.99 for observed versus expected values)

Low quality evidence from a survey of 10 clinical staff involved in a myeloma program (Gundlapalli et al, 2009) reported that clinic staff would be in favour of a single diagnostic report with the ability to view serial changes in key biomarkers and also supported the idea of providing a composite report directly to the patient.

Evidence Tables

Download PDF (743K)

Excluded Studies

Review Question

How should level of care be defined and categorised for people with haematological cancers who are having intensive (non-transplant) chemotherapy considering:

• Diagnosis
• Comorbidities and frailty
• Medicine Regimens
• Management of medicine administration and toxicities

Does the level of care affect patient outcome for people with haematological cancers who are having intensive, non-transplant chemotherapy, considering;

• Location
• Staffing levels
• Centre size/specialism
• Level of in-patient isolation
• Ambulatory care
• Prophylactic anti-infective medications

Background

Most patients who require curative treatment for aggressive haematological malignancies such as acute leukaemia or high-risk myelodysplastic syndrome, receive several cycles of intensive chemotherapy and the protocols used to treat these patients typically lower the blood cell count leading to severe neutropenia resulting in a neutrophil count of less than 0.5 ×10⁹/L. Other toxicities may also be a feature, and older patients and those with co-morbidities are at a higher risk of complications.

Despite improvements in supportive care, these patients are at a high risk of serious and potentially life-threatening infections and other complications.

In recent multicentre UK studies, early mortality following AML induction chemotherapy has been reported as up to 6% and 9% at 30 days and 10% and 15% at 60 days in younger and older patients respectively (Burnett et al Blood 2015; 125, 3878-3885, Burnett et al JCO 2012; 30,3924-31).

Reported induction mortality is also substantial in ALL; 4% in patients <55 and 18% in patients over 55 years (Sive et al BJH 2012;157:463-71). Early mortality in ALL is not improved with the introduction of modern drugs, such as tyrosine kinase inhibitors in Philadelphia positive disease (Fielding AK et al Blood 2014;123, 843-50). Recent data confirm a 2.2% induction death rate in 16-25 year olds treated on paediatric protocols. In 25 – 60 year olds treated on the current NCRI UKALL 14 type schedule, the induction death rate in UKALL 14 currently is 8.5% (personal communication, Dr Clare Rowntree).

Given the high risks of treatments and complexity of patients and speed complications can occur, immediate availability of specialist nursing staff supported initially by medical staff and then by prompt availability) of specialist staff (i.e. consultant/registrar) cover is essential, along with prompt access to other key specialists, especially intensive care. Specialist support services, especially specialist radiology and laboratory medicine (including transfusion medicine), are also essential on both an emergency and elective basis.

Along with adequate staffing and access to specialist services, the previous 2003 IOG recommended that patients treated on these protocols were nursed for the duration of their neutropenia (14-21 days) in specialist hospital units equipped in single rooms with or without laminar flow or high-efficiency particulate air (HEPA) filtration to reduce the risk of infection. Whereas this became common practice across many NHS units, for a variety of reasons, some patients receive care on an open ward or be allowed home, either through an informal arrangement with ward staff, or, increasingly through the structured delivery of intensive treatment in carefully selected patients (e.g. younger patients with limited co-morbidities) in the ambulatory care setting. However, promptness of clinical review by specialist staff also has to be in place for ambulatory care, where forward planning and policies are of major importance as the patient will have the additional ‘lag-phase’ of having to self-refer from home or hospital flat before assessment. This has to be balanced against NHS deliverability within working directives and generic/non-specialist hospital at night initiatives etc.

Despite being stipulated by the previous IOG and peer review recommendations, the provision of isolation rooms to protect intensively treated patients against nosocomial infections has proved challenging for NHS units despite rising levels of C. difficile, VRE, MRSA and other antibiotic resistant strains, along with seasonal respiratory viral infections (like influenza) in this population of patients, who are also susceptible to airborne fungal infections.. Although the benefits of isolation are well established in some contexts, it is not clear whether particular levels of protection are more effective in preventing fungal, bacterial and viral infections in severely immunocompromised patients than others. (e.g. standard en-suite rooms compared to more complex laminar flow and HEPA filtration). In any unit, isolation facilities, whether they are simple or complex, are a limited resource. Mandatory NHS isolation policies, designed to protect hospital inpatients as a whole, may impact significantly on bed availability for the intensively treated acute leukaemia patients, particularly during infectious epidemics such as influenza or outbreaks of antibiotic resistant infection. If isolation rooms for this patient population are not available at short notice, chemotherapy treatments may be delayed, or patients looked after in open wards or at home with informal arrangements, all of which may affect survival outcomes.

The standards of care required to deliver chemotherapy to patients with haematological cancer were previously classified according to the complexity of chemotherapy and duration of neutropenia. The 2003 haemato-oncology IOG and subsequent peer review standards stipulated a minimum of five intensive level 2 patients had to be treated per year but the recommendations were imprecise and open to interpretation, with both new and relapsed patients and a number of less intense lymphoma salvage regimens (such as DHAP and ESHAP etc) being potentially included. A further system of classification came from the updated BCSH recommendations. Three levels of care were defined predominantly relating to the facilities and support services required for patient care (BCSH Haematology Task Force, 2009). Whilst there was recognition that some patients may be managed from home, there was no major consideration of delivery of chemotherapy in the ambulatory care setting. Factors such as minimum numbers of patients required per ‘level’ of care, staff training and competency assessments were not specifically addressed in the BCSH guidelines for the facilities required for the treatment of adults with haematological malignancy (BCSH Haematology Task Force, 2009),

For haematopoietic stem cell transplantation (HSCT), the international FACT-JACIE accreditation standards for transplant programme stipulate minimum numbers for clinical activity. Despite early deaths from intensive induction chemotherapy for acute leukaemia being consistently higher than those associated with autologous stem cell transplantation (where UK adult 100 and 365 day non-relapse mortality is 2%) and closer to that for allogeneic transplantation (7% at 100 days rising to 16% at 1 year), there is no well defined minimum recommended threshold for unit activity in intensive chemotherapy (reference BSBMT 6^th Report to Specialist Commissioners (2015), http://BSBMT.org). Although minimum transplant activity thresholds are not evidence based, there is evidence that implementation of FACT-JACIE standards in haematological practice results in survival benefits for high-risk treatments (Gratwohl A et al Haematologica 2014; 99; 908-15).. There is also a case for having enough patients to perform meaningful analysis of survival outcomes and other audits within any unit undertaking intensive and complex treatments in this high-risk but potentially curable population of cancer patients.

In this IOG update there is a need to review and make clear evidence based recommendations for 24 hour specialist staffing levels and accessibility to isolation facilities, ITU and other support specialities. These are complex facilities and minimum numbers of patients with acute leukaemia and related conditions patients being treated with intensive chemotherapy in an individual unit need consideration in this IOG update. The update takes into account the potential clinical, patient experience and economic impact of intensive chemotherapy treatment in conventional or ambulatory care settings. Age and co-morbidities will also be a necessary consideration.

Levels of Care

A range of different levels of care, corresponding with the variety of diseases treated by haematology services, is required to manage patients with haematological cancers. Patients with acute leukaemia need repeated periods of intensive in-patient treatment lasting between four and seven months (depending on their diagnosis); 85-95% will be re-admitted as emergencies with febrile neutropenia on repeated occasions during this time (Flowers et al). By contrast, patients with conditions at the opposite end of the spectrum of aggressiveness, such as stage A chronic lymphocytic leukaemia, may need little more than regular monitoring.

The level of care required is based primarily on the duration and depth of neutropenia associated with different chemotherapy regimens. Patients being treated with regimens or dose schedules with a risk of brief and / or mild neutropenia can be managed on an outpatient basis. Patients being treated with regimens that usually cause prolonged, severe neutropenia, with a high risk of febrile neutropenia, require additional support and facilities. While some patients requiring these regimens may be treated in an outpatient setting, pathways need to be put in place to allow rapid access to inpatient care as required.

The British Committee for Standardisation in Haematology (BCSH) guidelines currently define four levels of care (level 1, 2a, 2b and 3). Level 2b is currently defined as treatment regimens which encompasses those that will predictably cause prolonged periods of neutropenia, would normally be given on an inpatient basis, and which may need to be given at weekends as well as during the week. According to BCSH guidelines, these regimens are more complex to administer than at the current level 1 or 2a and have a greater likelihood of resulting in medical complications in addition to predictable prolonged neutropenia. Consequently, the resources required to deliver these more complex regimens are greater than those needed for level 1 or 2a regimens. Level 3 care refers to complex regimens such as therapy for acute lymphoblastic lymphoma.

Historically, patients receiving treatment for Burkitt lymphoma or salvage chemotherapy for Hodgkin lymphoma and diffuse large B cell lymphoma were considered to be at risk of severe neutropenia. As a result these patients were treated according to the guidelines for level 2b patients. Data for the commonly used salvage regimens (e.g. DHAP, ESHAP and GDP with or without Rituximab) however show that these patients have a much lower risk of prolonged, severe neutropenia than previously thought. Consequently these patients may not require the same complex level of care, resource or facilities use as patients requiring induction therapy for conditions such as acute myeloid leukaemia or Burkitt lymphoma.

The guideline committee considered both the original levels of care defined in the NICE Haematology IOG (2003) and the two versions of the BCSH Guidelines (1995 and 2009) in conjunction with published data relating to toxicity of different regimens with the aim of redefining level 2b and 3 care from the BCSH guidelines and level 2 care from the IOG 2003, using a new definition based solely on the depth and duration of severe neutropenia expected for each regimen and patient group. The levels of care have therefore been redefined as non-intensive chemotherapy, intensive chemotherapy and haematopoietic Stem cell transplantation (HSCT, covering both autologous and allogeneic HSCT procedures).:

This guideline is concerned with patients receiving intensive chemotherapy regimens. The definition of intensive chemotherapy is any regimen which is anticipated to result in severe neutropenia of less than 0.5 ×10⁹/L for greater than 7 days, which largely limits the chemotherapy regimens to those used for AML (including acute promyelocytic leukaemia), high-risk MDS, ALL and Burkitt and lymphoblastic lymphomas (table 1).

Table 1

Levels of Care.

The use of other regimens that produce this degree of neutropenia is rare, but exceptional intensive treatment of other haematological malignancies is not excluded from this definition (table 2).

Table 2

Chemotherapy regimens and associated toxicities.

Question in PICO format

Searching and Screening

Total References retrieved (after databases combined, de-duplicated and sifted): 558

Medline search strategy (This search strategy is adapted to each database.)

1. exp Hematologic Neoplasms/
2. ((haematolog* or hematolog* or blood or red cell* or white cell* or lymph* or marrow or platelet*) adj1 (cancer* or neoplas* or oncolog* or malignan* or tumo?r* or carcinoma* or adenocarcinoma* or sarcoma*)).tw.
3. exp Lymphoma/
4. lymphoma*.tw.
5. (lymph* adj1 (cancer* or neopla* or oncolog* or malignan* or tumo?r*)).tw.
6. hodgkin*.tw.
7. lymphogranulomato*.tw.
8. exp Lymphoma, Non-Hodgkin/
9. (nonhodgkin* or non-hodgkin*).tw.
10. lymphosarcom*.tw.
11. reticulosarcom*.tw.
12. Burkitt Lymphoma/
13. (burkitt* adj (lymphom* or tumo?r* or cancer* or neoplas* or malign*)).tw.
14. brill-symmer*.tw.
15. Sezary Syndrome/
16. sezary.tw.
17. exp Leukemia/
18. (leuk?em* or AML or CLL or CML).tw.
19. exp Neoplasms, Plasma Cell/
20. myelom*.tw.
21. (myelo* adj (cancer* or neopla* or oncolog* or malignan* or tumo?r*)).tw.
22. kahler*.tw.
23. Plasmacytoma/
24. (plasm?cytom* or plasm?zytom*).tw.
25. (plasma cell* adj3 (cancer* or neoplas* or oncolog* or malignan* or tumo?r* or carcinoma* or adenocarcinoma*)).tw.
26. Waldenstrom Macroglobulinemia/
27. waldenstrom.tw.
28. exp Bone Marrow Diseases/
29. exp Anemia, Aplastic/
30. (aplast* adj an?em*).tw.
31. exp Myelodysplastic-Myeloproliferative Diseases/
32. exp Myeloproliferative Disorders/
33. exp Myelodysplastic Syndromes/
34. exp Thrombocytopenia/
35. (thrombocytop?eni* or thrombocyth?emi* or poly-cyth?emi* or polycyth?emi* or myelofibros or myelodysplas* or myeloproliferat* or dysmyelopoietic or haematopoetic or hematopoetic).tw.
36. exp Anemia, Refractory/
37. (refractory adj an?em*).tw.
38. (refractory adj cytop?en*).tw.
39. Monoclonal Gammopathy of Undetermined Significance/
40. (monoclonal adj gammopath*).tw.
41. (monoclonal adj immunoglobulin?emia).tw.
42. MGUS.tw.
43. ((oncohaematolog* or oncohematolog*) adj2 (disorder* or disease* or syndrome*)).tw.
44. or/1-42
45. limit 44 to yr=“2000 - 2015”
46. exp Antineoplastic Combined Chemotherapy Protocols/st
47. exp Antineoplastic Agents/st
48. Antimetabolites, Antineoplastic/st
49. (chemotherap* adj (regim* or protocol* or combin*)).tw.
50. intensive chemotherap*.tw.
51. (immunochemotherap* or immuno-chemotherap*).tw.
52. polychemotherap*.tw.
53. or/46-52
54. FLAG.tw.
55. Fludarabine/
56. Cytarabine/
57. Granulocyte Colony-Stimulating Factor/
58. 55 and 56 and 57
59. ((fludarabine or fludara) and (cytarabine or “Ara C” or “cytosine arabinoside”) and (g-csf or granulocyte colony-stimulating factor)).tw.
60. 54 or 58 or 59
61. FLAG-IDA.tw.
62. Idarubicin/
63. 58 and 62
64. (idarubicin or zavedos).tw.
65. 59 and 64
66. 61 or 63 or 65
67. DHAP.tw.
68. exp Dexamethasone/
69. Cisplatin/
70. 68 and 69 and 56
71. ((dexamethasone or decadron or oradexon or dexafree or dexsol) and (cytarabine or “Ara C” or “cytosine arabinoside”) and (cisplatin or platinol)).tw.
72. 67 or 70 or 71
73. ESHAP.tw.
74. Etoposide/
75. exp Methylprednisolone/
76. 74 and 75 and 56 and 69
77. ((etoposide or VP-16 or etopophos or vepesid) and (cytarabine or “Ara C” or “cytosine arabinoside”) and (cisplatin or platinol) and methylprednisolone).tw.
78. 73 or 76 or 77
79. IVE.tw.
80. Ifosfamide/
81. Epirubicin/
82. 80 and 81 and 74
83. ((ifosfamide or mitoxana) and (epirubicin or pharmorubicin) and (cytarabine or “Ara C” or “cytosine arabinoside”) and (etoposide or VP-16 or etopophos or vepesid)).tw.
84. 79 or 82 or 83
85. ICE.tw.
86. Carboplatin/
87. 80 and 86 and 74
88. ((ifosfamide or mitoxana) and carboplatin and (etoposide or VP-16 or etopophos or vepesid)).tw.
89. 85 or 87 or 88
90. (mini-BEAM or BEAM).tw.
91. Carmustine/
92. Melphalan/
93. 91 and 74 and 56 and 92
94. ((carmustine or BICNU) and (etoposide or VP-16 or etopophos or vepesid) and (cytarabine or “Ara C” or “cytosine arabinoside”) and melphalan).tw.
95. 90 or 93 or 94
96. DT-PACE.tw.
97. Thalidomide/
98. Doxorubicin/
99. Cyclosphamide/
100. 68 and 97 and 69 and 98 and 99 and 74
101. ((dexamethasone or decadron or oradexon or dexafree or dexsol) and (thalidomide or celgene) and (cisplatin or platinol) and (doxorubicin or adriamycin) and cyclophosphamide and (etoposide or VP-16 or etopophos or vepesid)).tw.
102. 96 or 100 or 101
103. CODOX-M IVAC.tw.
104. Vincristine/
105. Methotrexate/
106. 99 and 104 and 98 and 105 and 74 and 80 and 56
107. (cyclophosphamide and (vincristine or oncovin) and (doxorubicin or adriamycin) and methotrexate and (etoposide or VP-16 or etopophos or vepesid) and (ifosfamide or mitoxana) and (cytarabine or “Ara C” or “cytosine arabinoside”)).tw.
108. 103 or 106 or 107
109. DA.tw.
110. Daunorubicin/
111. 56 and 110
112. (daunorubicin and (cytarabine or “Ara C” or “cytosine arabinoside”)).tw.
113. 109 or 111 or 112
114. ADE.tw.
115. 56 and 110 and 74
116. ((cytarabine or “Ara C” or “cytosine arabinoside”) and daunorubicin and (etoposide or VP-16 or etopophos or vepesid)).tw.
117. 114 or 115 or 116
118. (FLAG or FLAG-IDA or DHAP or ESHAP or IVE or ICE or BEAM or mini-BEAM or DT-PACE or CODOX-M IVAC or DA or ADE).ps.
119. 53 or 60 or 66 or 72 or 78 or 84 or 89 or 95 or 102 or 108 or 113 or 117
120. rituximab.tw.
121. 119 and 120
122. 119 or 121
123. exp Precursor Cell Lymphoblastic Leukemia-Lymphoma/
124. leuk?emi*.tw.
125. (akut$ or acut$).tw.
126. 124 and 125
127. 123 or 126
128. Induction Chemotherapy/
129. Consolidation Chemotherapy/
130. (chemotherap* adj2 (induction or consolidat* or intensi*)).tw.
131. or/128-130
132. 127 and 131
133. 122 or 132
134. 45 and 133
135. exp Health Services/ma, st, ut
136. models, organizational/
137. exp Health Resources/og, st, ut
138. exp “Delivery of Health Care”/ma, mt, og, st, ut
139. Health Services Accessibility/og, st
140. Patient-Centered Care/ma, mt, og, st, ut
141. patient care plan*.tw.
142. Health Facilities/ma, st, ut
143. exp Health Facility Size/ma, og, st, sd
144. Health Manpower/
145. Specialization/
146. “Delivery of Health Care, Integrated”/
147. (“model* of care” or “level* of care” or “care model*” or “standard* of care” or “care standard*”).tw.
148. (“care coordination” or “care co-ordination”).tw.
149. (specialist* or expert* or expertise).tw.
150. Centralized Hospital Services/
151. ((integrat* adj3 healthcare) or (integrat* adj3 health care) or (integrat* adj3 service*) or (integrat* adj3 care*)).tw.
152. ((integrat$ adj3 provision) or (integrat$ adj3 organisation$)).tw.
153. (supercentre* or supercenter* or “super centre*” or “super center*”).tw.
154. exp Regional Health Planning/
155. ((local adj hospital*) or facility* or centre* or center* or service* or clinic* or unit* or site*).tw.
156. ((outreach or satellite*) adj (healthcare or health care or care or service* or centre* or center* or clinic* or unit* or department* or facilit* or site*)).tw.
157. co-locat*.tw.
158. Cancer Care Facilities/
159. Oncology Service, Hospital/
160. Medical Oncology/ma, og, st
161. Ancillary Services, Hospital/
162. (support* adj (service* or facilit* or unit* or department* or on-site)).tw.
163. ((haematolog* or hematolog* or haemato-oncolog* or hemato-oncolog* or oncolog*) adj2 (service* or facilit* or unit* or department* or on-site)).tw.
164. outpatients/
165. ambulatory care facilities/
166. exp Ambulatory Care/ma, st, ut
167. (ambulatory care or ambulatory health care or ambulatory healthcare).tw.
168. (ambulatory service* or ambulatory health service*).tw.
169. Outpatient Clinics, Hospital/
170. (outpatient* or out-patient*).tw.
171. Day Care/ma, og, st, ut
172. (day adj (care or case* or unit* or facilit*)).tw.
173. Hospital Shared Services/
174. shared care.tw.
175. exp Hospitalization/
176. ((hospital* or inpatient* or in-patient* or patient*) adj (admission* or admitted or readmission* or re-admission* or readmitted or re-admitted)).tw.
177. Patient Isolation/
178. (patient* adj2 isolat*).tw.
179. Hemodialysis Units, Hospital/
180. exp Emergency Medical Services/ma, og, st, ut
181. (emergenc* adj (healthcare or health care or care or service* or centre* or center* or clinic* or unit* or department* or facilit* or site*)).tw.
182. Intensive Care Units/
183. exp Critical Care/ma, og, st, ut
184. (critical care or intensive care or high dependency or ICU or HDU).tw.
185. (intensive therapy unit or ITU).tw.
186. exp health personnel/
187. staff*.tw.
188. (haematologist* or hematologist* or haemato-oncologist* or hemato-oncologist* or oncologist*).tw.
189. Nursing Services/
190. Oncology Nursing/
191. (nurs* adj2 (haematolog* or hematolog* or haemato-oncolog* or hemato-oncolog*)).tw.
192. Nurse's Role/
193. Clinical Nursing Research/
194. Inservice Training/og, st
195. Pharmacies/ma, og, st, ut
196. exp Pharmaceutical Services/
197. Pharmacists/
198. exp Home Care Services/
199. (home adj2 (care or nursing or service*)).tw.
200. exp Community Health Services/
201. (communit* adj2 (care or nursing or service* or clinic*1 or unit* or centre* or center*)).tw.
202. Social Support/
203. Palliative Care/ma, og, st, ut
204. Catheterization, Central Venous/st, ut
205. (prophyla* adj2 (anti-fungal* or antiviral* or antibiotic*)).tw.
206. Catheter-Related Infections/pc
207. Bacterial Infections/pc
208. Bacteremia/pc
209. Cross Infection/pc
210. exp Infection Control/mt, og, st
211. Environment, Controlled/
212. *Filtration/
213. HEPA filtration.tw.
214. high efficiency particulate air filtration.tw.
215. (air adj2 (filtration or filter*)).tw.
216. or/135-215
217. 134 and 216

Screening Results

Study Quality

The evidence for this topic comprises one systematic review and meta-analysis; one randomised trial; one randomised cross-over study; one prospective study; one audit and four retrospective comparative studies.

A number of factors were identified which impacted the quality of the evidence including study populations which were not exclusively low risk haematology patients, retrospective, non-randomised methodology, selection bias, small sample sizes and possible recall bias.

Evidence Statements

Isolation Factors

Survival

Very low to moderate quality evidence (Grade table 1) from one systematic review and meta-analysis which included 40 studies (randomised trials and observational) (Schlesinger et al, 2009); protective isolation with any combination of methods that included air quality control reduced the risk of death at 30 days (RR=0.6; 95% CI 0.5-0.72; 15 studies, 6280 patients); 100 days (RR=0.79, 95% CI, 0.73-0.87; 24 studies, 6892 patients) and at the longest available follow-up (between 100 days and 3 years) (RR=0.86, 95% CI 0.81-0.91; 13 studies, 6073 patients).

Grade Table 1

Isolation compared to No isolation/Placebo for low risk patients.

Infection related Mortality, Risk of Infection, Antibiotic use

Very low to moderate quality evidence (Grade table 1) from one systematic review and meta-analysis which included 40 studies (randomised trials and observational) (Schlesinger et al, 2009); protective isolation reduced the occurrence of clinically and/or microbiologically documented infections (RR=0.75 (0.68-0.83) per patient; 20 studies, 1904 patients; RR=0.53 (0.45-0.63); per patient day, 14 studies, 66431 patient days).

Very low to moderate quality evidence (Grade table 1) from one systematic review and meta-analysis which included 40 studies (randomised trials and observational) (Schlesinger et al, 2009); no significant benefit of protective isolation (all studies used air quality control) was observed in relation to mould infections (RR=0.69, 0.31-1.53; 9 studies, 979 patients) nor was the need for systemic antifungal treatment reduced (RR=1.02, 95% CI 0.88-1.18; 7 studies, 987 patients).

Very low to moderate quality evidence (Grade table 1) from one systematic review and meta-analysis which included 40 studies (randomised trials and observational) (Schlesinger et al, 2009); gram positive and gram negative infections were significantly reduced, though barrier isolation was needed to show a reduction in gram negative infections (RR= 0.49 (0.40-0.62) with barrier isolation (12 trials/n=1136) versus RR=0.87 (0.61-1.24) without barrier isolation (4 trials/n=328).

Very low to moderate quality evidence (Grade table 1) from one systematic review and meta-analysis which included 40 studies (randomised trials and observational) (Schlesinger et al, 2009); the need for systemic antibiotics did not differ when assessed on a per patient basis (RR=1.01, 0.94-1.09; 5 studies, 955 patients) but the number of antibiotic days was significantly lower with protective isolation (RR=0.81, 0.78-0.85; 3 studies, 6617 patient days).

Room facilities

Very low quality evidence from one retrospective cohort-control study (Grade table 1) comparing outcomes before and after ward renovation in 63 patients (Hutter et al, 2009) reported that patients treated before renovation (2 patients per room, 6 patients sharing a toilet placed outside the room, wash basin inside the room, shower across the hospital corridor, no ventilation system, air filtration or room pressurisation, no false ceilings) stayed 3 days longer compared with those treated on the newly renovated ward (2 patients per room, separate bath room in each room equipped with toilet, wash basin and shower, no ventilation system, air filtration or room pressurisation, no false ceilings). 39% of pre-renovation patients and 34% of post-renovation patients developed an invasive pulmonary aspergillus (p=0.79) with the diagnosis usually determined on CT scan.

Ambulatory Care

Survival

Very low quality evidence (grade table 2) from one systematic review and meta-analysis (Schlesinger et al, 2009); febrile patients were discharged for further antibiotic treatment at home if stable. All cause mortality was significantly lower in the outpatient setting (RR=0.72, 95% CI 0.53-0.97) at longest follow-up (median follow-up 12 months; range 1-36).

Grade Table 2

Ambulatory Care versus inpatient care.

Unpublished data collected by the Sheffield Ambulatory Care Unit and University College Hospital, London Ambulatory Care Unit reported no deaths in the Ambulatory Care Unit between during the period January 2011-March 2015 (Appendix 1).

Hospital Admissions and length of stay

Very low quality evidence (grade table 2) from one UK audit of a hotel based, ambulatory care unit (Sive et al, 2012b); there were 1443 admissions to the Ambulatory Care Unit (9126 patient days) during the study period(688 patients from 18-79 years of age), whose length of stay ranged from 1 to 42 days (median 5). 82% of admissions were in haematology oncology patients with lymphoma being the largest single group of patients by days of use. Patients receiving less myelosuppressive regimens tended to be discharged home on treatment completion while patients receiving more intensive treatment almost always required readmission to the ward at some point. 813/1443 (56%) patients were discharged directly home; 53/630 (9%) patients admitted to the ward were scheduled in advance

Very low quality evidence (grade table 2) from one UK audit of a hotel based, ambulatory care unit (Sive et al, 2012b), 456/576 (79%) of unscheduled ward admissions were within ACU working hours, 66 (11%) were out of hours and 54 (9%) had no time recorded. The most common reason for unscheduled admission included infection or fever, nausea and vomiting and poor oral intake or dehydration.

Very low quality evidence (grade table 2) from one retrospective study in which patients who were fit for home care were given a choice between home care and inpatient care (Sopko et al, 2012); 17/41 patients required ambulatory management only while 24 patients required re-hospitalisation, primarily due to febrile neutropenia. In 36 febrile episodes a microbiologically documented infection was the most common cause of fever (61%) with the remaining episodes being of unknown origin. Patients re-hospitalised were admitted for a mean 10.9 days (6-35 days) versus a mean hospitalisation time of 30 days for inpatients (17-38). Mean duration of hospitalisation for inpatients from the time they became febrile to discharge was 14.3 days (7-22 days).

Very low quality evidence (grade table 2) from one retrospective analysis of 30 patients (Bakshi et al, 2009); 25/69 consolidation cycles resulted in hospital admission and all were associated with febrile neutropenic episodes or documented infections. Hospital stay was significantly shorter in outpatient cycles compared with inpatient cycles (p<0.001) leading to a saving of 269 patient-days for the entire study group.

Unpublished data collected by the Sheffield Ambulatory Care Unit and University College Hospital, London Ambulatory Care Unit was combined to calculate inpatient bed days saved through the use of an ambulatory care program. An average of sixteen inpatient bed days per patient was saved for Acute Myeloid Leukaemia, an average of nine inpatient bed days were saved for Acute Lymphoblastic Leukaemia and sixteen inpatient bed days for Burkitt Lymphoma (Appendix 1)

Infections

Very low quality evidence (grade table 2) from one systematic review and meta-analysis (Schlesinger et al, 2009); febrile patients were discharged for further antibiotic treatment at home if stable and febrile neutropenia or documented infections occurred less often in the outpatient group (RR=0.78, 95% CI 0.7-0.88; 8 studies, 757 patients), rates of bacteraemia were lower in the outpatient group but the difference was not significant (RR=0.68, 95% CI 0.43-1.05; 2 studies. 252 patients).

Very low quality evidence (grade table 2) from one retrospective analysis of 30 patients (Bakshi et al, 2009); significantly fewer outpatients required second line antibiotics compared with inpatients (p=0.03) and mean duration of antibiotic administration was significantly lower in the outpatient group (p=0.04).

Transfusions

Very low quality evidence (grade table 2) from one retrospective analysis of 30 patients (Bakshi et al, 2009); a median of 1 (0-4) unit of packed red blood cells was transfused per consolidation cycle in the outpatient setting and 2 (0-5) in the inpatient setting and a median of 1 (0-13) platelet transfusions were administered at the outpatient clinic and 2 (0-12) in the inpatient setting.

Quality of Life

Very low quality evidence (grade table 2) from one randomised cross over trial (Stevens et al, 2005) quality of life for 29 paediatric patients treated at home or in hospital (standard care) was assessed, children in the home group experienced a decrease in factor 1 (sensitivity to restrictions in physical functioning and ability of maintain a normal physical routine) of the POQOLS measures when they switched from home based treatment to hospital based treatment with an average change of 5.2 while standard care patients experienced an improvement in QoL when they switched to home based treatment with an average score of -10.5 (p=0.023)

Patients in the home-based group had significantly higher scores for factor 2 (emotional distress) measures compared with the hospital treatment group (pair wise comparison at the end of each 6 months phase p=0.043).

Very low quality evidence (grade table 2) from a nested qualitative study (Stevens et al, 2004) within a randomised cross over trial (Stevens et al, 2005); 33 health practitioners (hospital and community based) reported that home-based care seemed to have a positive impact on daily life and psychological well-being of children and families particularly in relation to disruption and psychological stress, reporting a reduction in disruption due to reduced travelling, reduced hospital clinic waiting time and reduced time missed from school and work.

“I think the big advantage is certainly it helps the children and their families to maintain a more normal routine on that day – to be able to avoid having to miss work and school – and have a big disruption and cost added to their day to come all the way down here for treatment that could be provided in a much shorter period and at a time that's more convenient for them.”

Health practitioners also reported noting fewer signs of psychological distress in children and parents during the home chemotherapy phase; children appeared happier and more comfortable while parents appeared to have more of a sense of control over the illness and treatment.

“Most kids seem to like it [chemotherapy] at home; they are happier. But I find that with community nursing in general. Some of the kids are so withdrawn when they come into the hospital, and are so different at home. So are the parents. Parents are usually more at ease at home, feel they have more control at home.”

The advantages conferred by consistency in personnel and practice were emphasised by hospital based practitioners. Children in the hospital setting were seen by the same practitioner helping parents and children become comfortable and trusting while in the community setting, care providers were less consistent.

“I'm the consistent person that gives the chemotherapy and the children; they adapt to you and the way you do things, and you get to know them. That's consistent, that helps them.” [Clinic Nurse]

“Whoever was working that day would go to see the patients. It was mostly the three of us…whoever was working was going. It took longer, but generally not in the first time but within a few times; they would get comfortable with the procedure” [Community Nurse]

Patient Satisfaction

Very low quality evidence (grade table 2) from one retrospective study in which 17 patients were treated at home for 46 cycles (Luthi et al, 2012); patients reported that they were ‘very satisfied’ with home care and one case reported being ‘satisfied’. None of the patients showed a preference for inpatient care for the next chemotherapy cycles. 38% of patients stated a preference for home care and others had no declared preference. Patient reported benefits of home care included a higher comfort level (100%), freedom and possibility to organise their own time (94%) and the reassurances and comfort of having a relative present (88%). 78% of patients were not concerned about the absence of a nurse and87% did not record any anxiety during home care treatment

Very low quality evidence (grade table 2) from one retrospective study in which 17 patients were treated at home for 46 cycles (Luthi et al, 2012), the main patient reported disadvantages were feelings of dependency on a relative (19%) and or being a burden (6%) however, relatives who returned questionnaires (63%) and all were in favour of home care and 97% were in favour of home care for next treatment.

Primary concerns about home care included the presence of strangers (nurse, physician) at home (16%), request for continuous presence as patients were not allowed to be alone for more than one hour (14%), anxiety and fatigue (14%) and lack of freedom for leisure and holidays (14%).

Burden of Care

Very low quality evidence (grade table 2) from one randomised cross over trial (Stevens et al, 2005) including 29 paediatric patients treated at home or in hospital (standard care) reported no evidence of an effect of the location of chemotherapy administration was observed on the parental burden of care (assessed using the care giving burden scale).

Impact on Practitioners

Very low quality evidence (grade table 2) from a nested qualitative study (Stevens et al, 2004) within a randomised cross over trial (Stevens et al, 2005)suggested that community health practitioners should have specific education in relation to home care, administration of chemotherapy to children and meeting psychological needs of children with cancer and their families. Four home care nurses took part in a three day educational session on chemotherapy administration and reported that they found the course extremely valuable.

Very low quality evidence (grade table 2) from a nested qualitative study (Stevens et al, 2004) within a randomised cross over trial (Stevens et al, 2005); health practitioners agreed that the major benefit of hospital treatment was that the resources and treatments were all centralised and coordinated.

“Their [children and parents] only experience has been with [hospital name] and you whip your child in and they get a little finger poke and then sometimes an hour or two later the results are back and then it's very smooth.”

While having home chemotherapy, children had to go to community laboratories to have their blood tests carried out, many technicians lacked paediatric experience and were insensitive to their needs.

“The biggest one [problem] we have run into has been the whole lab issue and the fact that we've discovered that laboratories in the community are not very child friendly [hospital programme director]

There was also an issue with laboratory results not being communicated to the community nurses for subsequent drug prescription and home delivery resulting in increased workload while nurses retrieving results from hospital physicians. It was suggested that there should be one central person to liaise between the hospital and community.

Very low quality evidence (grade table 2) from a nested qualitative study (Stevens et al, 2004) within a randomised cross over trial (Stevens et al, 2005); some hospital physicians reported feeling less confident about prescribing chemotherapy agents for children due to the inability to assess the child directly and be in charge of the healthcare process in the community. They also reported feeling unclear about issues relating to liability and responsibility.

Very low quality evidence (grade table 2) from a nested qualitative study (Stevens et al, 2004) within a randomised cross over trial (Stevens et al, 2005); 2 clinic nurses and 3 paediatric oncologists reported no change in their workload; 5 clinic nurses and 1 physician reported an increase due to the increased volume of paperwork and 3 clinic nurses reported a decrease. 13/14 community health practitioners reported an increase in workload primarily due to increased paperwork and increased time communicating with other health practitioners to expedite the process.

“It has added to my responsibilities, the day before having to give chemo, I am doing a lot of phone calling. Labs, clinic, chemo. it can be very time consuming and very frustrating but the actual visit time is not the issue.” [community nurse]

Community practitioners reported they had increased their repertoire of skills and ‘felt good’ about helping families which increased their personal satisfaction. It was also reported that partnership between community and hospital was enhanced by effective communication with opportunities to collaborate and share ideas and optimise treatments.

Very low quality evidence (grade table 2) from a nested qualitative study (Stevens et al, 2004) within a randomised cross over trial (Stevens et al, 2005); the home chemotherapy programme was associated with less interaction with children and families which was considered to be both a positive (fewer patients in outpatient clinics, health practitioners less busy, more time for children in attendance) and negative (distressing because they were not sure how the children were coping with treatment) thing.

“You look forward to their visits, I do anyways. Because the communication of how they're really doing and how things are going is sort of broken down, there's a gap because you don't see them every two weeks.” [hospital clinic nurse]

Responses suggested an increased level of frustration as the home chemotherapy programme was challenging to accommodate in terms of scheduling between health practitioners and families.

“I found that we were juggling a lot. Trying to work around the teenagers schedules because you would end up calling them to say that you were going to come and do the chemo and they would say ‘Oh no I'm off to something or other tonight’ So I had to go the home early at 7:30 the next morning. So of course we tried to do that but when you have a lot of patients you just cannot do it. We can't always work around their schedule and I think that really needs to be made clear.” [community nurse]

Feasibility

Very low quality evidence (grade table 2) from one retrospective study in which 17 patients were treated at home for 46 cycles (Luthi et al, 2012); home treatment required 1 physician visit and 2 nurse visits per day accounted for 621 visits during 46 treatment cycles (207 days of home treatment). 32 additional home visits were required as a result of technical problems with the pump (median, 1 visit per cycle; range 0-4 visits per cycle) and most visits were needed at the start of treatment.

Pump failure due to air bubbles was the main technical problem and was resolved by flushing the tube (n=21 cases).

Partial disconnection at the exit channel occurred in 9 cases and needle disconnection from the port of the catheter occurred in 2 cases

2 major pump failures were reported resulting in one overnight hospitalisation and a 4 day hospitalisation.

Advice on restrictions on social contact, pets and food

From one retrospective audit of 336 institutions in 27 countries (Lehrnbecher et al, 2012), 107 centres (32%) had written protocols for non-pharmacological anti-infective approaches and n=64 (64%) had a general agreement without a written policy. In 85 centres (25%) practitioners used an individualised approach

A physician was involved in the instruction of parents in 89% (n=299) of centres and a nurse in 71% of centres (n=238).

A handout was provided to parents in 52% (n=174) of centres and was the only information given in 4% (n=14) of cases.

42% of parents received a handout and were additionally provided with verbal information by a nurse or physician.

Restriction scores in Europe were significantly higher than in USA, suggesting greater restrictions; restriction scores did not differ by centre.

In relation to social contact, most centres do not allow children with AML to visit indoor public places, attend daycare , nursery or school while recommendations for patients with ALL varied considerably. Restrictions mostly related to neutropenia (58%) and to chemotherapy regimens and the health of surrounding people was a pre-condition for reduced restrictions in 16% of centres.

In relation to pets, there was wide variation in recommendations for both AML and ALL patients. Restrictions under certain circumstances related to appropriate hand-washing after contact (27%), keeping animals already at home without introducing new pets (25%), restriction of pets in the bedroom or on the bed(22%), ensuring pets were assessed by a veterinary specialist (17%) and restrictions on cleaning of cages/litter trays (16%).

In relation to food, most centres had restrictions on raw meat, raw seafood and unpasteurised milk for both AML and ALL patients. There were wide variations in food restrictions around salad, nuts, takeaway food and unpeeled vegetables. In 68% of cases, restrictions were generally related to neutropenia and specific chemotherapy regimens. If uncooked vegetables or salad were allowed, appropriate cleaning was advised (12%).

In relation to the use of facemasks, 9% (n=30) institutions recommended children with ALL wear face masks in public while 34% (n=114) recommended face masks for AML patients. 54% (n=181) never suggest facemasks for children with ALL and 41% (n=138) never suggest facemasks for children with AML.

Evidence Tables

Download PDF (545K)

Appendix 1. Ambulatory Care Data

Ambulatory Care Data provided by UCHL (personal communication Barbara von Barsewisch) and Sheffield (personal communication John Snowdon)

Excluded Studies