Health Economics & Outcome Research: Open Access

A Ten-Step Model for Planning and Costing Environmental Health Services in Healthcare Facilities

Ben Martin^{* *}Correspondence: Ben Martin, Health Economics and Outcome Research, Brussels, Belgium, Email:

Author info »

Abstract

To offer quality patient care and avoid healthcare-acquired illnesses, a clean atmosphere is crucial. Understanding prices is critical for planning service delivery budgets, but there is little cost evidence for Environmental Health Services (EHS) in Healthcare Facilities (HCFs). The first comprehensive evaluation of the costs of setting up, running, and maintaining EHS in HCFs in Low- And Middle-Income Countries is presented (LMICs). Water, sanitation, hygiene, cleaning, waste management, personal protective equipment, vector control, laundry, and lighting costs in LMICs were thoroughly looked for. Our search turned up 36 studies with expenses for 51 EHS Water expenses were reported in three studies, sanitation costs in three, hygiene costs in four, waste management costs in thirteen, cleaning costs in sixteen, personal protective equipment costs in two, laundry costs in ten, and lighting costs in none. The evidence was of poor quality. The reported expenses were rarely indicative of the entire costs of providing EHS. Costs per unit were rarely published. This review identifies opportunities to improve costing research through efforts to categorise and disaggregate EHS costs, increased dissemination of previously unpublished data, improvements to indicators to monitor EHS demand and quality required to contextualise costs, and the development of frameworks to define EHS needs and essential inputs to guide future costing.

Introduction

To offer quality patient care and avoid healthcare-acquired illnesses, a safe, sanitary Healthcare Environment is critical (HAIs). In high-income nations, HAIs constitute the major cause of avoidable morbidity and death among hospitalised patients. Despite the lack of data from low- and middle-income countries (LMICs), the burden of HAIs is anticipated to be higher where environmental health services (EHS) are lacking .According to systematic evaluations of HAIs in Africa and LMICs across the world around 15% of hospitalised patients get a HAI, and one in every 17 patients with a HAI dies from associated reasons . HAIs are also linked to longer hospital stays, resulting in billions of dollars in medical costs that could have been avoided. In order to provide safe treatment in healthcare institutions, it is necessary to maintain a sanitary atmosphere (HCFs). Inadequate environmental conditions can limit the likelihood of seeking medical help and increase the risk of HAIs . In Low- And Middle-Income Countries (LMICs), HAIs are estimated to afflict 15% of all hospitalised patients and are the primary cause of mortality among hospitalised patients [6]. HAIs are caused by dangerous environmental conditions and poor hygiene among healthcare workers in 60%-80% of cases. Environmental Health Services (EHS) at Healthcare Facilities (HCFs) are crucial for ensuring a safe, functional environment, but their costs remain unknown. Progress toward universal access to EHS in HCFs is hampered by a lack of awareness of costs. In a network of medical research and training institutions in Lilongwe, Malawi, we defined frameworks for key costs required to offer EHS and conducted an expost financial analysis of EHS in seven outpatient buildings, servicing an estimated 42,000 patients yearly. Water, sanitation, hygiene, usage of personal protective equipment at the point of care, waste management, cleaning, washing, and vector control were all projected to be expensive. We utilised worldwide rules and standards to construct frameworks of key outputs and inputs for each EHS, which we used to identify expenses required for EHS delivery and evaluate the completeness of costs data in our case study. We adopt a mixed-methods approach to pricing, including qualitative interviews to gain a better understanding of the facility's environment and reviewing computerised data to calculate expenses. We assessed the initial expenses of setting up services as well as the yearly expenditures of operations and maintenance.

Search Update Strategy

During the search upgrade, we applied machine learning to help with title and abstract screening. Machine learning algorithms can leverage the results of manual screening in the original search as training data for machine learning algorithms to automate aspects of the screening process for studies given in search updates, significantly lowering screening time and effort To select studies from the search update for manual screening using the doctor programme (Document Classification and Topic Extraction Resource), we employed semi-supervised learning and supervised machine learning in two rounds (ICF, Virginia, USA). All of the research The results of the search update that were manually inspected were examined in accordance with our protocol's recommendations (see Supplementary Information) Doctor uses the language of titles and abstracts to prioritize search results and includes clustering, supervised clustering, and supervised machine learning techniques. For the first phase of prioritizing, we employed supervised clustering with an ensemble technique.

Supervised clustering is a type of semi-supervised learning in which a collection of known relevant (i.e. "seed") studies is used to categories an undifferentiated corpus of research. Seed studies are assumed to have a high proportion of relevant research, hence clusters with a high number of seed studies are prioritized for manual screening. Great length into supervised clustering and show that it can compete with supervised machine learning methods in terms of accuracy while using less training data. The ensemble method employs three cluster sizes: k-means and Nonnegative Matrix Factorization (NMF). Using each method with the three different cluster numbers results in six different clustering models (for example, the KM-10 model is a k-means algorithm with 10 clusters, while the KM-20 model is a kmeans algorithm with 20 clusters).

The six models were applied to the title and abstract text of a collection of seed studies in a literature search update. Seed studies are a type of training data, however they need fewer positive trials than machine learning algorithms generally require. The 154 relevant studies found through title and abstract screening in the original search were included in our seed set. The result of supervised clustering using a six-model ensemble technique is an ensemble score for each research ranging from 0 to 6, indicating the number of models in which the study was identified in a relevant cluster (i.e., a cluster with a high proportion of seed studies).

Studies having an ensemble score of 3 or above were carefully screened. The remaining studies were passed to supervised machine learning for a second round of prioritising. Training data were taken from the previously screened research, and studies with a probability of relevance of 0.5 or above were manually vetted. The result of supervised clustering using a six-model ensemble technique is an ensemble score for each research ranging from 0 to 6, indicating the number of models in which the study was identified in a relevant cluster (i.e., a cluster with a high proportion of seed studies).

Studies having an ensemble score of 3 or above were carefully screened. The remaining studies were passed to supervised machine learning for a second round of prioritising. Training data were taken from the previously screened research, and studies with a probability of relevance of 0.5 or above were manually vetted.

Theoretical Foundation

Social theories establish a more extensive relationship between health and wellbeing factors and processes. The article is based on political ecology of health theory, which investigates how power, politics, institutions, agendas, and/or agents impact the environment and population health hazards. This theory goes on to look at how global health discourse influences and shapes local settings such as policy creation and implementation. At the global, national, and local levels, the prioritising, execution, and management of WaSH interventions are political and power-laden. This idea has proved beneficial in the study of development project prioritisation and implementation as well as health.

It has also influenced research on healthcare services in LMICs and the effects of water privatisation on health and well-being.