Systematic review
Task sharing for selected health services in hospitals
Systematic review
|Updated
The purpose of this systematic review was to summarize the effects of task sharing for some selected procedures in hospitals.
Key message
The purpose of this systematic review was to summarize the effects of task sharing for some selected procedures in hospitals. For four of the six procedures, we found no evidence that met our criteria for inclusion. Based on evidence assessed as being from low to very low quality we draw the following conclusions for two of six questions: What are the effects of task sharing between doctors and nurses for patients undergoing endoscopy? ● There may not be large differences for patient outcomes such as: pain/discomfort, gastrointestinal symptoms and quality of life. For the outcomes: need of assistance, duration, number of polyps missed, depth of sigmoidoscopy, number of biopsies, immediate complications and costs we cannot, on the basis of the evidence, determine whether there are important differences between endoscopy performed by nurses or by doctors. What are the effects of task sharing between doctors and nurses for patients followed up in outpatient clinics? ● Patients with bronchiectasis: there may not be large differences in quality of life or the number of hospitalisations per patient. However, it is possible that there may be greater costs associated with the use of nurses. For the outcomes: lung function, lung capacity, exacerbations due to infection and training capacity we cannot, on the basis of the evidence, determine whether there are important differences for patients between follow-up by nurses or by doctors. ● Patients with asthma: there may not be large differences in quality of life, hospitalisation per patient or costs. For outcomes: number of symptom-free days, lung function, number of medication-free days, maximum air flow or the number of exacerbations we cannot, on the basis of the evidence, decide whether there are important differences for patients between follow-up by nurses or by doctors. ● Patients with rheumatoid arthritis: patient satisfaction may be somewhat improved due to follow-up by nurses rather than rheumatologists. For general health outcomes: joint pain, fatigue, global assessment of disease activity and disease activity measured by DAS 28 it is possible that there are no large differences between follow-up by nurses or by rheumatologists. ● Adults with cancer: nurses may use more time on consultation and take more blood tests, but there may not be large differences for patient satisfaction if cancer patients are followed up by a doctor or nurse. For the outcomes: mental health, depression, occurrence of metastases or overall costs we cannot, on the basis of the evidence, decide whether there are important differences for patients between follow-up by nurses or doctors.
Summary
Background "Improved task sharing between staff in hospitals" was one of the Minister of Health and Care Services ten proposed measures to improve hospital function. For some health services in hospitals, capacity can be less than the demand. This may result in long queues and unwanted latency for patients. Consequences are delayed diagnosis and treatment, increased length of stay in hospital and increased pressure on health workers. By transferring certain tasks from one occupational group working in a field with great pressure, to another cadre, the idea is to reduce queuing. Those who traditionally performed the task are relieved and free to perform other tasks. The purpose of this systematic review was to summarise the research that has evaluated the effects of task sharing between health workers in the specialist health care level for selected health services and assess the quality of the evidence of the results. We pre-specified the following six areas for task sharing: relieve doctors in endoscopy, relieve doctors in monitoring patients in outpatient clinics in general, relieve the operating room nurses, relieve radiologists, relieve pathologists, and to share tasks from medical staff to medical secretaries and other mercantile personnel. Methods We searched for systematic reviews and randomized controlled trials in several relevant medical databases and the ISI Web of Science in June 2013. Titles and abstracts were reviewed, potentially relevant publications ordered in full text and assessed for inclusion or exclusion. We included reports that addressed at least one of our six questions. Only systematic reviews of high quality, assessed with the Norwegian Knowledge Centre for the Health Services checklist, and randomized controlled trials were included. The results were assessed for risk of bias. The quality of the overall evidence for each outcome, were assessed using GRADE. For all phases of the study selection and critical reviews two people were involved and upon disagreement, a third person was consulted. Results Four systematic reviews and five randomized controlled trials met our inclusion criteria. They addressed task sharing in two of the six questions we wanted to study. Three primary studies compared endoscopy performed by nurses with endoscopy performed by doctors. In four systematic reviews and in two primary studies doctors were compared to nurses for outpatient follow-up and monitoring of patients with bronchiectasis, asthma, cancer or rheumatoid arthritis. We identified no relevant studies addressing the other four questions about the effects of transferring tasks from the operating room nurses to surgical technicians, from radiologists to radiographers, from pathologists to pathologist assistants, and effects of new roles for medical secretaries. For endoscopy performed by nurses compared with endoscopy performed by physicians the evidence suggests that patients may not experience significant differences for the outcomes: pain/discomfort, gastrointestinal symptoms and quality of life. For the outcomes: need of assistance, time, number of polyps missed, depth of sigmoidoscopy, number of biopsies, immediate complications and costs, evidence is not provided to determine whether there are important differences between endoscopy performed by nurses or doctors. For outpatient follow-up by nurses compared with physicians for patients with bronchiectasis or asthma, evidence suggests that there may not be large differences in quality of life or the number of hospitalisations per patient. However, it is possible that there may be greater costs when patients with bronchiectasis are followed-up by nurses, while costs of follow-up of patients with asthma by nurses - versus physicians may not be very different. For the outcomes: lung function, lung capacity, maximum air flow, number of symptom-free days, number of medication-free days, exacerbations or exercise capacity, evidence is not provided to determine whether there are important differences between using specialist nurses or physicians for follow-up of patients with bronchiectasis or asthma. For follow-up of adults with rheumatoid arthritis, evidence suggests that patient satisfaction may be better when followed-up by nurses rather than rheumatologists. For general health outcomes: joint pain, fatigue, global assessment of disease activity and disease activity measured by DAS 28, there may not be large differences between follow-up by nurses or by rheumatologists. For follow-up of adults with cancer, evidence suggests that nurses may spend more time on consultation and take more blood tests than doctors. Furthermore, there may not be a large difference in patient satisfaction if the patient is follow-up by a nurse or doctor. For the outcomes: incidence of metastases (spread), general mental health, depression, or overall costs, evidence is not provided to determine whether there are important differences between follow-up by nurses or physicians. Discussion We found fragmented and inadequate research-based evidence about effects of task sharing within the six areas of specialist health care that we had pre-specified. However, it is possible that we may have missed relevant studies because the questions required a complicated literature search. The studies we summarised could only partially answer two of our six questions. Broadly speaking, the authors of the included studies concluded that there were no differences between interventions if the results were not statistically significantly different. However, when a result is not statistically significant one must examine the size of the estimated uncertainty around the power estimate; the confidence interval. Then you can see both the lowest possible and the highest possible value the effect estimate can assume and one must consider whether the possible differences can then be so large that they become clinically important. If so, one cannot conclude that there is no difference. We have tried to give a more balanced and accurate picture of the results than the authors in the primary studies had done. Conclusion We identified research evidence that partially answered the two questions about the effects of task sharing in endoscopy and follow-up of patients in outpatient clinics. For the other questions, we identified no research evidence. The available evidence suggests that there may not be large differences for some outcomes in task-shifting for nurses and doctors in both endoscopy and follow-up of patients in outpatient clinics. The evidence for several other important outcomes is insufficient, which means that we cannot determine whether there may be significant differences in health care or costs between using nurses or doctors. A key issue concerning task sharing is patient safety; will new ways of organising tasks among health professions affect patient safety? Especially, will there be differences in prevalence of rare events, such as the incidence of perforation by sigmoidoscopy? In such cases, one must have a large amount of patient data to get a sufficient number of events to detect any potential difference. Wide confidence intervals indicate that the studies have not included a sufficient number of participants. To answer the question about task sharing and patient safety, there is a need for more studies with sufficient sample size to find the minimum difference that could be reasonable to exclude an important difference or not - i.e. equivalence studies or non-inferiority studies.