Personalising intravenous to oral antibiotic switch decision making through fair interpretable machine learning.

Antimicrobial resistance (AMR) and healthcare associated infections pose a significant threat globally. One key prevention strategy is to follow antimicrobial stewardship practices, in particular, to maximise targeted oral therapy and reduce the use of indwelling vascular devices for intravenous (IV) administration. Appreciating when an individual patient can switch from IV to oral antibiotic treatment is often non-trivial and not standardised. To tackle this problem we created a machine learning model to predict when a patient could switch based on routinely collected clinical parameters. 10,362 unique intensive care unit stays were extracted and two informative feature sets identified. Our best model achieved a mean AUROC of 0.80 (SD 0.01) on the hold-out set while not being biased to individuals protected characteristics. Interpretability methodologies were employed to create clinically useful visual explanations. In summary, our model provides individualised, fair, and interpretable predictions for when a patient could switch from IV-to-oral antibiotic treatment. Prospectively evaluation of safety and efficacy is needed before such technology can be applied clinically.

Virtual reality technology for surgical learning: qualitative outcomes of the first virtual reality training course for emergency and essential surgery delivered by a UK-Uganda partnership.

INTRODUCTION: The extensive resources needed to train surgeons and maintain skill levels in low-income and middle-income countries (LMICs) are limited and confined to urban settings. Surgical education of remote/rural doctors is, therefore, paramount. Virtual reality (VR) has the potential to disseminate surgical knowledge and skill development at low costs. This study presents the outcomes of the first VR-enhanced surgical training course, 'Global Virtual Reality in Medicine and Surgery', developed through UK-Ugandan collaborations. METHODS: A mixed-method approach (survey and semistructured interviews) evaluated the clinical impact and barriers of VR-enhanced training. Course content focused on essential skills relevant to Uganda (general surgery, obstetrics, trauma); delivered through: (1) hands-on cadaveric training in Brighton (scholarships for LMIC doctors) filmed in 360°; (2) virtual training in Kampala (live-stream via low-cost headsets combined with smartphones) and (3) remote virtual training (live-stream via smartphone/laptop/headset). RESULTS: High numbers of scholarship applicants (n=130); registrants (Kampala n=80; remote n=1680); and attendees (Kampala n=79; remote n=556, 25 countries), demonstrates widespread appetite for VR-enhanced surgical education. Qualitative analysis identified three key themes: clinical education and skill development limitations in East Africa; the potential of VR to address some of these via 360° visualisation enabling a 'knowing as seeing' mechanism; unresolved challenges regarding accessibility and acceptability. CONCLUSION: Outcomes from our first global VR-enhanced essential surgical training course demonstrating dissemination of surgical skills resources in an LMIC context where such opportunities are scarce. The benefits identified included environmental improvements, cross-cultural knowledge sharing, scalability and connectivity. Our process of programme design demonstrates that collaboration across high-income and LMICs is vital to provide locally relevant training. Our data add to growing evidence of extended reality technologies transforming surgery, although several barriers remain. We have successfully demonstrated that VR can be used to upscale postgraduate surgical education, affirming its potential in healthcare capacity building throughout Africa, Europe and beyond.

Effect of neurosurgical high observation beds (HOBs) on service provision: a single-centre experience and cost analysis.

OBJECTIVE: To evaluate the effect of level 1, high observation beds (HOBs) compared with high dependency unit (HDU) and neurosurgical intensive care unit (NICU) admission on service provision, such as cancelled operations, and healthcare costs. METHODS: A retrospective, observational, single-centre cross-sectional study at a single, large UK neurosurgical centre. All adult patients admitted to neurosurgical HOBs between December 2021 and July 2022 were included. The list of cancelled procedures was collected from 2019 to 2022. To evaluate the impact of admission of eligible patients to HOBs, the total bed days, cost per bed day, number of admissions and cost per admission were obtained for all clinical areas the financial years 2019/2020 and 2021/2022. RESULTS: 307 patients were included in the study: 59.7% of HOBs admissions were elective and 37.7% were acute; admissions were for cranial procedures or conservative treatment (64.8%), spinal (32.6%) or other (2.6%). Following admission, 73.3% of patients were stepped down to the ward prior to discharge home. Only seven patients required escalation to level-2 or level-3 care. Overall, 97% of all HOBs patients were discharged home at the end of hospital stay. Occupancy rate was 90.4%.The cost of bed day increased from ward, level 0 (£384), then level 1 (£376), to level 2 (£787-1211) and to level 3 (£1628). From 2019 to 2021, 558 operations had been cancelled, and 140 (37.8%) of 370 were estimated to have been potentially avoided by HOBs admissions due to conflict of scheduling, ward bed capacity and critical care bed capacity. In addition, a minimum total expenditure due to cancelled operations was estimated at £22 923.50 yearly on average. CONCLUSION: This study recognises HOBs growing role in the management of acutely unwell patients in ward-based environments. While recognising the associated challenges, this study highlighted the potential in reducing healthcare costs. Further studies should evaluate the impact and limitations of HOBs on patients' recovery and outcomes, compared to HDU and NICU.

Utilising an accelerated Delphi process to develop consensus on the requirement and components of a pre-procedural core robotic surgery curriculum.

Robot-assisted surgery (RAS) continues to grow globally. Despite this, in the UK and Ireland, it is estimated that over 70% of surgical trainees across all specialities have no access to robot-assisted surgical training (RAST). This study aimed to provide educational stakeholders guidance on a pre-procedural core robotic surgery curriculum (PPCRC) from the perspective of the end user; the surgical trainee. The study was conducted in four Phases: P1: a steering group was formed to review current literature and summarise the evidence, P2: Pan-Specialty Trainee Panel Virtual Classroom Discussion, P3: Accelerated Delphi Process and P4: Formulation of Recommendations. Forty-three surgeons in training representing all surgical specialties and training levels contributed to the three round Delphi process. Additions to the second- and third-round surveys were formulated based on the answers and comments from previous rounds. Consensus opinion was defined as ≥ 80% agreement. There was 100% response from all three rounds. The resulting formulated guidance showed good internal consistency, with a Cronbach alpha of > 0.8. There was 97.7% agreement that a standardised PPCRC would be advantageous to training and that, independent of speciality, there should be a common approach (95.5% agreement). Consensus was reached in multiple areas: 1. Experience and Exposure, 2. Access and context, 3. Curriculum Components, 4 Target Groups and Delivery, 5. Objective Metrics, Benchmarking and Assessment. Using the Delphi methodology, we achieved multispecialty consensus among trainees to develop and reach content validation for the requirements and components of a PPCRC. This guidance will benefit from further validation following implementation.

There is a great variety of orthopaedic conditions presenting to a large hospital in Sierra Leone: a 3-year prospective study.

PURPOSE: In Sierra Leone there is a large void in orthopaedic research into the type of orthopaedic injuries, both acute and chronic. Improved data collection is essential in providing insight to guide health care planning and research. This study aims to outline the types of orthopaedic injury sustained. METHOD: Data were prospectively collected by local surgeons in the Orthopaedic outpatient department at a large hospital between January 2016 and January 2019. RESULTS: The orthopaedic department saw a mean 728 patients per year, with mean age 24.0 years. The workload comprised of 64.92% acute orthopaedic conditions or their complications, with 35.08% elective orthopaedics. Fractures made up the largest proportion of clinical appointments, annually 244.33 fractures; however there was a high incidence of osteomyelitis. CONCLUSION: The study gives an important insight into the types and distribution of elective and trauma orthopaedic injuries sustained in Sierra Leone, which has not been previously reported, and highlights key areas where resources may be focused in order to improve clinical outcomes.

Machine learning and synthetic outcome estimation for individualised antimicrobial cessation.

The decision on when it is appropriate to stop antimicrobial treatment in an individual patient is complex and under-researched. Ceasing too early can drive treatment failure, while excessive treatment risks adverse events. Under- and over-treatment can promote the development of antimicrobial resistance (AMR). We extracted routinely collected electronic health record data from the MIMIC-IV database for 18,988 patients (22,845 unique stays) who received intravenous antibiotic treatment during an intensive care unit (ICU) admission. A model was developed that utilises a recurrent neural network autoencoder and a synthetic control-based approach to estimate patients' ICU length of stay (LOS) and mortality outcomes for any given day, under the alternative scenarios of if they were to stop vs. continue antibiotic treatment. Control days where our model should reproduce labels demonstrated minimal difference for both stopping and continuing scenarios indicating estimations are reliable (LOS results of 0.24 and 0.42 days mean delta, 1.93 and 3.76 root mean squared error, respectively). Meanwhile, impact days where we assess the potential effect of the unobserved scenario showed that stopping antibiotic therapy earlier had a statistically significant shorter LOS (mean reduction 2.71 days, p -value <0.01). No impact on mortality was observed. In summary, we have developed a model to reliably estimate patient outcomes under the contrasting scenarios of stopping or continuing antibiotic treatment. Retrospective results are in line with previous clinical studies that demonstrate shorter antibiotic treatment durations are often non-inferior. With additional development into a clinical decision support system, this could be used to support individualised antimicrobial cessation decision-making, reduce the excessive use of antibiotics, and address the problem of AMR.

Day 2 neutrophil-to-lymphocyte and platelet-to-lymphocyte ratios for prediction of delayed cerebral ischemia in subarachnoid hemorrhage.

OBJECTIVE: Recent evidence has suggested that an admission neutrophil-to-lymphocyte ratio (NLR) of ≥ 5.9 predicts delayed cerebral ischemia (DCI) in aneurysmal subarachnoid hemorrhage (aSAH). The primary aims of this study were to assess reproducibility and to ascertain the predictive ability of NLR on subsequent days postictus. Secondary aims included identification of additional inflammatory markers. METHODS: A single-center, retrospective study of all patients aged ≥ 18 years with aSAH between May 2014 and July 2018 was performed. Patient characteristics, DCI incidence, operative features, and outcomes (on discharge and at 3 months postictus) were recorded. C-reactive protein (CRP) and full blood count differentials were recorded on admission and through day 8 postictus or at discharge. In total, 403 patients were included in the final analysis. RESULTS: Ninety-six patients (23.8%) developed DCI with a median time from ictus of 6 days (IQR 3.25-8 days). A platelet-to-lymphocyte ratio (PLR) cutoff ≥ 157 and CRP cutoff ≥ 27 was used in our cohort. In a multiple binary logistic regression model, after controlling for known DCI predictors, day 2 NLR ≥ 5.9 (OR 2.194, 95% CI 1.099-4.372; p = 0.026), day 1 PLR ≥ 157 (OR 2.398, 95% CI 1.1072-5.361; p = 0.033), day 2 PLR ≥ 157 (OR 2.676, 95% CI 1.344-5.329; p = 0.005), and CRP ≥ 27 on days 3, 4, and 5 were predictive of DCI. CONCLUSIONS: The results of this study have confirmed the association between NLR and DCI and have demonstrated the predictive potential of PLR and CRP, suggesting that NLR and PLR at day 2, and CRP from day 3 onward, may be better predictors of DCI than those measurements at the time of ictus.

Impact of surgical drain output monitoring on patient outcomes in hepatopancreaticobiliary surgery: A systematic review.

BACKGROUND AND OBJECTIVE: Surgical drains are widely utilized in hepatopancreaticobiliary surgery to prevent intra-abdominal collections and identify postoperative complications. Surgical drain monitoring ranges from simple-output measurements to specific analysis for constituents such as amylase. This systematic review aimed to determine whether surgical drain monitoring can detect postoperative complications and impact on patient outcomes. METHODS: A systematic review was performed, and the following databases searched between 02/03/20 and 26/04/20: MEDLINE, EMBASE, The Cochrane Library, and Clinicaltrials.gov. All studies describing surgical drain monitoring of output and content in adult patients undergoing hepatopancreaticobiliary surgery were considered. Other invasive methods of intra-abdominal sampling were excluded. RESULTS: The search returned 403 articles. Following abstract review, 390 were excluded and 13 articles were included for full review. The studies were classified according to speciality and featured 11 pancreatic surgery and 2 hepatobiliary surgery studies with a total sample of 3262 patients. Postoperative monitoring of drain amylase detected pancreatic fistula formation and drain bilirubin testing facilitated bile leak detection. Both methods enabled early drain removal. Improved patient outcomes were observed through decreased incidence of postoperative complications (pancreatic fistulas, intra-abdominal infections, and surgical-site infections), length of stay, and mortality rate. Isolated monitoring of drain output did not confer any clinical benefits. CONCLUSIONS: Surgical drain monitoring has advantages in the postoperative care for selected patients undergoing hepatopancreaticobiliary surgery. Enhanced surgical drain monitoring involving the testing of drain amylase and bilirubin improves the detection of complications in the immediate postoperative period.

Barriers to accessing and providing rehabilitation after a lower limb amputation in Sierra Leone - a multidisciplinary patient and service provider perspective.

PURPOSE: The primary aim was to explore the perceived barriers that lower limb amputees and service providers face when accessing or providing rehabilitation services. The secondary aim was to describe the lower limb amputations performed in public hospitals in the Western Area of Sierra Leone in 2018. MATERIALS AND METHODS: A mixed methodology was employed, involving the collection of amputation data from surgical logbooks and interviews with amputees (n = 10) and group discussion and interviews with service providers (n = 11). RESULTS: Of the 37 primary lower limb amputations (49% men, 51% women; median age 56 years; 62% transtibial and 35% transfemoral amputations) 86% were for diabetic and vascular causes. Barriers to accessing services included poor transportation access, high service fees, rural living, gender and a lack of government support. Insufficient funding and supplies, skilled staff shortages and a lack of local training programmes were frequently reported barriers to providing rehabilitation services. CONCLUSIONS: A low prioritisation means rehabilitation services are underfunded, resulting in numerous barriers to both accessing and providing amputee rehabilitation services. Subsidised services and an outreach programme may improve access for patients. Increased funding and local training programmes are needed to improve service delivery.Implications for RehabilitationComprehensive and accessible amputee rehabilitation services can enable people with amputations to regain their independence and aid their participation in their community and workplace.There are numerous barriers to both accessing and providing amputee rehabilitation services in the Western Area, Sierra Leone, chiefly financial. We recommend a revised effort by the Sierra Leonean government to implement the progressive policies on disability they have already adopted into law, which will aid the improvement of amputee rehabilitation services. New education and training programmes for all levels of prosthetic and orthotic professions are needed to secure the future of prosthetics and orthotics in Sierra Leone.

Chronic wounds in Sierra Leone: Searching for Buruli ulcer, a NTD caused by Mycobacterium ulcerans, at Masanga Hospital.

BACKGROUND: Chronic wounds pose a significant healthcare burden in low- and middle-income countries. Buruli ulcer (BU), caused by Mycobacterium ulcerans infection, causes wounds with high morbidity and financial burden. Although highly endemic in West and Central Africa, the presence of BU in Sierra Leone is not well described. This study aimed to confirm or exclude BU in suspected cases of chronic wounds presenting to Masanga Hospital, Sierra Leone. METHODOLOGY: Demographics, baseline clinical data, and quality of life scores were collected from patients with wounds suspected to be BU. Wound tissue samples were acquired and transported to the Swiss Tropical and Public Health Institute, Switzerland, for analysis to detect Mycobacterium ulcerans using qPCR, microscopic smear examination, and histopathology, as per World Health Organization (WHO) recommendations. FINDINGS: Twenty-one participants with wounds suspected to be BU were enrolled over 4-weeks (Feb-March 2019). Participants were predominantly young working males (62% male, 38% female, mean 35yrs, 90% employed in an occupation or as a student) with large, single, ulcerating wounds (mean diameter 9.4cm, 86% single wound) exclusively of the lower limbs (60% foot, 40% lower leg) present for a mean 15 months. The majority reported frequent exposure to water outdoors (76%). Self-reports of over-the-counter antibiotic use prior to presentation was high (81%), as was history of trauma (38%) and surgical interventions prior to enrolment (48%). Regarding laboratory investigation, all samples were negative for BU by microscopy, histopathology, and qPCR. Histopathology analysis revealed heavy bacterial load in many of the samples. The study had excellent participant recruitment, however follow-up proved difficult. CONCLUSIONS: BU was not confirmed as a cause of chronic ulceration in our cohort of suspected cases, as judged by laboratory analysis according to WHO standards. This does not exclude the presence of BU in the region, and the definitive cause of these treatment-resistance chronic wounds is uncertain.

What factors shape surgical access in West Africa? A qualitative study exploring patient and provider experiences of managing injuries in Sierra Leone.

INTRODUCTION: Surgical access is central to universalising health coverage, yet 5 billion people lack timely access to safe surgical services. Surgical need is particularly acute in post conflict settings like Sierra Leone. There is limited understanding of the barriers and opportunities at the service delivery and community levels. Focusing on fractures and wound care which constitute an enormous disease burden in Sierra Leone as a proxy for general surgical need, we examine provider and patient perceived factors impeding or facilitating surgical care in the post-Ebola context of a weakened health system. METHODS: Across Western Area Urban (Freetown), Bo and Tonkolili districts, 60 participants were involved in 38 semistructured interviews and 22 participants in 5 focus group discussions. Respondents included surgical providers, district-level policy-makers, traditional healers and patients. Data were thematically analysed, combining deductive and inductive techniques to generate codes. RESULTS: Interacting demand-side and supply-side issues affected user access to surgical services. On the demand side, high cost of care at medical facilities combined with the affordability and convenient mode of payment to the traditional health practitioners hindered access to the medical facilities. On the supply side, capacity shortages and staff motivation were challenges at facilities. Problems were compounded by patients' delaying care mainly spurred by sociocultural beliefs in traditional practice and economic factors, thereby impeding early intervention for patients with surgical need. In the absence of formal support services, the onus of first aid and frontline trauma care is borne by lay citizens. CONCLUSION: Within a resource-constrained context, supply-side strengthening need accompanying by demand-side measures involving community and traditional actors. On the supply side, non-specialists could be effectively utilised in surgical delivery. Existing human resource capacity can be enhanced through better incentives for non-physicians. Traditional provider networks can be deployed for community outreach. Developing a lay responder system for first-aid and front-line support could be a useful mechanism for prompt clinical intervention.

The Incidence of Low Anterior Resection Syndrome as Assessed in an International Randomized Controlled Trial (MRC/NIHR ROLARR).

OBJECTIVE: To investigate the incidence of LARS in patients undergoing elective anterior resection within the MRC/NIHR ROLARR trial and to explore perioperative variables that might be associated with major LARS. SUMMARY BACKGROUND DATA: Sphincter-preserving rectal cancer surgery is frequently accompanied by defaecatory dysfunction known as Low anterior resection syndrome (LARS). This is distressing for patients and is an unmet clinical challenge. METHODS: An international, retrospective cohort study of patients undergoing anterior resection within the ROLARR trial was undertaken. Trial participants with restoration of gastrointestinal continuity and free from disease recurrence completed the validated LARS questionnaire between August 2015 and April 2017. The primary outcome was the incidence of LARS and secondary outcome was severity (minor versus major). RESULTS: LARS questionnaires were received from 132/155 (85%) eligible patients. The median time from surgery to LARS assessment was 1065 days (range 174-1655 d). The incidence of LARS was 82.6% (n = 109/132), which was minor in 26/132 (19.7%) and major in 83/132 (62.9%). The most common symptoms were incontinence to flatus (n = 86/132; 65.2%) and defaecatory clustering (88/132; 66.7%). In a multivariate model, predictors of major LARS were: 1 cm decrease in tumor height above the anal verge (OR = 1.290, 95% CI: 1.101,1.511); and an ASA grade greater than 1 (OR = 2.920, 95% CI: 1.239, 6.883). Treatment allocation (laparoscopic vs robotic) did not predict major LARS. CONCLUSIONS: LARS is a common after rectal cancer surgery and patients should be appropriately counselled preoperatively, particularly before surgery for low tumors or in comorbid populations.

Delayed Gastric Emptying Post-Esophagectomy: A Single-Institution Experience.

OBJECTIVE: Delayed gastric emptying (DGE) is a common functional disorder after esophagectomy in patients with esophageal carcinoma. Management of DGE varies widely and it is unclear how comorbidities influence the postoperative course. This study sought to determine factors that influence postoperative DGE. METHODS: This retrospective study evaluates patients who underwent esophagectomy with gastric pull-up between 2007 and 2019. The cohort was stratified in various ways to determine if postoperative care and outcomes differed, including patient demographics, comorbidities, intraoperative and postoperative procedures. RESULTS: During the study period, 149 patients underwent esophagectomy and 37 had diabetes. Overall incidence of DGE, as defined in this study, was 76.5%. Surgery type was significantly different between DGE and normal emptying cohorts (P = 0.005). Comparing diabetic and nondiabetic patients, there was no significant difference noted in DGE (P = 0.25). Additionally, there was no difference in presence of DGE for patients who underwent any intraoperative pyloric procedure compared to those who did not (P = 0.36). Of significance, all 16 patients with chronic obstructive pulmonary disease had a delay in gastric emptying (P = 0.01). CONCLUSIONS: A higher proportion of patients with DGE post-esophagectomy were identified compared to the literature. There is little consensus on a true definition of DGE, but we believe this definition identifies patients suffering in the immediate postoperative period and in follow-up. There is no evidence to support a different postoperative course for patients with diabetes, but the link between chronic obstructive pulmonary disease and DGE warrants further investigation.

Global academic response to COVID-19: Cross-sectional study.

This study explores the response to COVID-19 from investigators, editors, and publishers and seeks to define challenges during the early stages of the pandemic. A cross-sectional bibliometric review of COVID-19 literature was undertaken between 1 November 2019 and 24 March 2020, along with a comparative review of Middle East respiratory syndrome (MERS) literature. Investigator responsiveness was assessed by measuring the volume and type of research published. Editorial responsiveness was assessed by measuring the submission-to-acceptance time and availability of original data. Publisher-responsiveness was assessed by measuring the acceptance-to-publication time and the provision of open access. Three hundred and ninety-eight of 2,835 COVID-19 and 55 of 1,513 MERS search results were eligible. Most COVID-19 studies were clinical reports (n = 242; 60.8%). The submission-to-acceptance [median: 5 days (IQR: 3-11) versus 71.5 days (38-106); P < .001] and acceptance-to-publication [median: 5 days (IQR: 2-8) versus 22.5 days (4-48·5-; P < .001] times were strikingly shorter for COVID-19. Almost all COVID-19 (n = 396; 99.5%) and MERS (n = 55; 100%) studies were open-access. Data sharing was infrequent, with original data available for 104 (26.1%) COVID-19 and 10 (18.2%) MERS studies (P = .203). The early academic response was characterized by investigators aiming to define the disease. Studies were made rapidly and openly available. Only one-in-four were published alongside original data, which is a key target for improvement. Key points: COVID-19 publications show rapid response from investigators, specifically aiming to define the disease.Median time between submission and acceptance of COVID-19 articles is 5 days demonstrating rapid decision-making compared with the median of 71.5 days for MERS articles.Median time from acceptance to publication of COVID-19 articles is 5 days, confirming the ability to introduce rapid increases at times of crisis, such as during the SARS outbreak.The majority of both COVID-19 and MERS articles are available open-access.

Safety and efficiency of gasless laparoscopy: a systematic review protocol.

BACKGROUND: Gasless laparoscopy, developed in the early 1990s, was a means to minimize the clinical and financial challenges of pneumoperitoneum and general anaesthesia. It has been used in a variety of procedures such as in general surgery and gynecology procedures including diagnostic laparoscopy. There has been increasing evidence of the utility of gasless laparoscopy in resource limited settings where diagnostic imaging is not available. In addition, it may help save costs for hospitals. The aim of this study is to conduct a systematic review of the available evidence surrounding the safety and efficiency of gasless laparoscopy compared to conventional laparoscopy and open techniques and to analyze the benefits that gasless laparoscopy has for low resource setting hospitals. METHODS: This protocol is developed by following the Preferred Reporting Items for Systematic review and Meta-Analysis-Protocols (PRISMA-P). The PRISMA statement guidelines and flowchart will be used to conduct the study itself. MEDLINE (Ovid), Embase, Web of Science, Cochrane Central, and Global Index Medicus (WHO) will be searched and the National Institutes of Health Clinical Trials database. The articles that will be found will be pooled into Covidence article manager software where all the records will be screened for eligibility and duplicates removed. A data extraction spreadsheet will be developed based on variables of interest set a priori. Reviewers will then screen all included studies based on the eligibility criteria. The GRADE tool will be used to assess the quality of the studies and the risk of bias in all the studies will be assessed using the Cochrane Risk assessment tool. The RoB II tool will assed the risk of bias in randomized control studies and the ROBINS I will be used for the non-randomized studies. DISCUSSION: This study will be a comprehensive review on all published articles found using this search strategy on the safety and efficiency of the use of gasless laparoscopy. The systematic review outcomes will include safety and efficiency of gasless laparoscopy compared to the use of conventional laparoscopy or laparotomy. TRIAL REGISTRATION: The study has been registered in PROSPERO under registration number: CRD42017078338.