BOHR International Journal of Engineering

2023, Vol. 2, No. 1, pp. 1–6

DOI: 10.54646/bije.2023.11
www.bohrpub.com

Hospital recommender system

Shrey Sumariya * , Shreyas Rami * , Shubham Revadekar * , Vidhan Shah* and Sudhir Bagul *
Department of Computer Engineering, Dwarkadas J. Sanghvi College of Engineering, Mumbai, India

*Correspondence:
Shrey Sumariya,
[email protected]
Shreyas Rami,
[email protected]
Shubham Revadekar,
[email protected]
Vidhan Shah,
[email protected]
Sudhir Bagul,
[email protected]

Received: 05 January 2023; Accepted: 11 January 2023; Published: 11 January 2023

Elderly patients require more medical effort. It is clear that early-stage disease diagnosis can support timely
and appropriate treatment. But if you don’t pay attention in a timely manner, it can lead to different kinds of
health problems that can lead to death. Take advantage of our recommendation system to recommend hospitals.
A recommender system uses algorithms to provide product or service recommendations to users. By combining
blockchain technology and machine learning models, we provide users with highly accurate recommendations.
This whitepaper describes how sophisticated machine learning models and blockchain can be connected to
improve recommendations, providing hospitals with higher performance and more accurate recommendations.
An optimized model for recommending hospitals in a better manner is the main goal behind this paper.
Keywords: machine learning, smart contracts, recommendation system, TFID Vectorizer, cosine similarity

Introduction unreliable results. Health and medicine are one such field.
A recommendation system should not only avoid mistakes
Typical questions that traditional recommendation systems and support decision-making but should also understand
answer is about what kind of books are worth reading, patients, attitudes, requirements, and values in the context of
which movies are worth watching, and what products disease and health care. This makes health recommendation
are worth buying. A recommender system helps you systems more difficult to apply. Considering various diseases,
hospitals or doctor recommendation has many aspects to
sift through large amounts of data to determine the
consider before recommending someone. Cost, patient risk,
options that are the most relevant tasks that users have
location, disease type, doctor availability, and many other
in mind. Recommendation systems are everywhere when
factors should be considered when making a decision.
customers buy products from online stores, e.g., Amazon. Mistakes or errors in considering any of the above factors
Use an intelligent recommendation system that recommends could have serious consequences. Medical recommendations
different products using a content-based collaborative play an important role, but the risk factors are high, and
approach. Say something like, “Customers who buy this many aspects must be considered. Hospital referrals for
will buy that,” or put it another way, “There are some cancer patients are always a tedious task. There is a lot
products that might be useful, based on your searches here.” of information to consider when recommending a hospital.
Decisions made by customers are low risk, but decisions First of all, cancer is a high-risk disease, and in many cases,
in other areas often have more serious consequences, patients need to receive proper medical care, which requires
and recommendations based on those decisions may yield proper medical care in all aspects.

1
2 Sumariya et al.

In today’s world, information run and privacy area units the effect of institutional predictors. This can be calculated by
are therefore vital, and we have a tendency to tend to use experts to assess the outcomes of their problems.
blockchain for this purpose. Additionally, to provide privacy Physicians assign a case to a particular hospital focusing
protection, we have a tendency to produce a hash layer on the patients’ physical condition and distance to travel.
that is troublesome to hack to forestall information loss. An experienced physician can select the hospital that poses
A blockchain may be distributed info that keeps a history of the least risk to the patient. A patient is more likely to
all transactions that have ever occurred during a blockchain accept such a tailored approach. The goal of this project is to
network. All network nodes have a duplicate of this info, create an expert system that can help with such a personally
replicated and secured. The advantage of blockchain is that tailored hospital selection decision. Two significant priorities
untrusted participants will communicate and exchange assets are travel distance and survival likelihood. When the hospital
in a secure manner while not the necessity of a sure third selection decision is tailored, the tradeoff between different
party. As the name suggests, a blockchain may be a series aims can be handled explicitly. Due to the small result space,
of blocks joined along by a cryptological hash. Every block the problem is first framed as single-objective optimization
references the previous block by storing a cryptological in this study and could be addressed by a systematic search
hash, manufacturing a series of blocks. Additionally, to (only the number of hospitals). In this optimization, a query
the previous block’s hash, a block contains a collection provides the system with data about the patient, such as age,
of transactions that, once accepted and connected to the admission type, comorbidities, and the maximum suffered
blockchain, cannot be updated or deleted. As a result, each distance. To generate a personalized hospital selection, the
consistency and double-spending problem area unit was optimization process will combine the given information
alleviated. A wise contract may be a worm that runs on with the captured knowledge.
the blockchain. It is viewed as a sure third party between The goal of the gathering and analysis tool is to find the
leery participants. It consists of contract memory, credits, hospital with the highest likelihood of survival considering
and program code. By merely submitting dealing to the the greatest suffered distance to a hospital. If we also want
blockchain, it is created and utilized by any node on the to consider other factors in the healthcare decision, such
network. Sensible contracts’ code is fastened and cannot be as the likelihood of difficulties, we have a multi-objective
updated once it is on the blockchain. During this article, optimization problem.
I primarily wish to stipulate the associate in the nursing From (2), our strategy requires that there will be
application I needed to make for hospital referrals for cancer two different kinds of independent variables in the
patients. During the latter half, I will be able to make a problem. Uncontrollable (unchangeable) variables are the
case for everything we have a tendency to use, together first category. These variables’ values are fixed and cannot
with, however, our recommendation system considers several be changed. Patient characteristics such as demographic
factors in recommending hospitals. information, the results of diagnostic tests, the kind of
admission, whether surgery was performed, the patient’s
comorbidity scores, and the method of payment, for instance,
Literature review are uncontrollable variables in this study. Variables with
controlled (adjustable) values are the opposite type. These
In (1), the paper “Building a hospital referral expert system factors can be used to base the recommendation. The values
with a Prediction and Optimization-Based Decision Support of these variables in our application describe a hospital. We
System algorithm” proposes to design a recommendation suggest a new functionality for using blockchain technology
system for hospitals using artificial intelligence. Choosing to manage and clean the dataset. Based on the information
a healthcare facility with a history of providing quality provided by the blockchain, the controllable or modifiable
healthcare can make the distinction between asked issues variables in the dataset are frequently updated.
and incorrect issues, including death. Endured croakers From (3), the blockchain era is redefining facts of modeling
learn which facilities offer stylish care. Physicians generally and governance deployed in lots of healthcare applications.
play a crucial part in recommending specific hospitals for This is in particular because of its adaptability and ability to
their cases. Some medical situations are time-critical, and segment, steady, and proportion clinical facts and offerings
transportation time plays a veritably important part in issues. in an unheard-of way. The blockchain era is at the center of
Some medical situations require immediate attention, and many modern-day traits withinside the healthcare enterprise.
transportation time is critical. Distance is frequently the The blockchain era sits at the pinnacle of the uncooked
most important consideration when choosing a hospital for facts layer this is taken into consideration as the middle
patients living in pastoral and underserved areas. Even in framework in pursuit to create a secured healthcare structure
non-emergency circumstances, proximity is preferable. As this is divided into 4 components. Blockchain technology
a result, several research studies have shown that patients gets more and more effective and robust, as they come
frequently favor local higher-risk healthcare facilities to go to to be coupled with AI in numerous real-phrase healthcare
lower-risk health facilities. Geographic factors may influence solutions. Machine gaining knowledge and deep gaining
10.54646/bije.2023.11 3

knowledge is the primary using elements for the AI domain, fitness records may be effective, it is not without demanding
which is likewise relentlessly enhancing the development situations. Personalizing fitness records manually is time-
of automation. The extra facts we feed to the device and consuming and highly priced, and the outputs are regularly
the extra cap potential a device will benefit to categorize inconsistent. Natural language technology (NLG) strategies
or expect styles accurately. The blockchain era is gaining can address those issues and are consequently more and
sizable interest from individuals, in addition to groups more used withinside the fitness domain. The main instance
of almost a wide variety and dimensions. It is able to of tailor-made fitness records with the use of NLG is the
rework the conventional enterprise with its features, which BabyTalk device evolved through Gatt et al. (2009).
encompass decentralization, anonymity, persistence, and From (5–9), Ethereum is a public, open-source,
audibility. The blockchain era is anticipated to reshape the blockchain-based distributed computing platform and
healthcare ecosystem. Not best will the method be obvious operating system with smart contract functionality. Using
and steady, but additionally the exceptional of healthcare can blockchain, we are developing a decentralized and self-
be multiplied at a decreased cost. Specifically, we supplied tallying Internet voting protocol with maximum voter
modern-day studies on fitness fact control and the way privacy. The healthcare industry is one of the largest in the
blockchain will empower sufferers and streamline the sharing world, accounting for more than 10% of the gross domestic
method of fitness facts. We located that there is a consensus product (GDP) of the most developed countries. Blockchain
among researchers that, with the blockchain era, affected and its smart contract capabilities, in particular, have the
person facts can be surely owned and managed through the potential to address healthcare interoperability issues such
rightful proprietor of the facts, i.e., the affected person. The as enabling effective interactions between users and medical
blockchain permits fitness data to be time-stamped in order applications, securely delivering patient data to a wide range
that nobody can tamper with them after turning them into a of organizations and devices, and improving the overall
part of the distributor ledger. efficiency of medical practice workflow. Accessing and
The sufferers can have the proper to determine who can managing a large amount of medical data is also part of the
and cannot get the right of entry to their facts and for what job. The associated cost for this system has been estimated
purpose. However, there are nonetheless numerous open in terms of a feasibility study as part of the implementation
demanding situations that require additional investigation. of the workflows of the medical smart contract system
For instance, cross-border sharing of fitness facts in which for healthcare management. Not all customized NLG
distinct and regularly conflicting jurisdictions exist may also information is effective. Reiter et al. (2003) developed
avoid the advantage of blockchain’s facts sharing. In fact, “STOP,” a system for creating personalized smoking
the expectancy of a character’s privateness varies from one cessation letters. Nontailored letters, in contrast, were just as
United States of America to any other primarily based totally effective as tailored letters. We do not need to change health
on the authority’s regulations. Therefore, destiny studies on behavior in our system, but rather inform patients in the best
regulation, standardization, and cross-border fitness facts way possible to help them make decisions (4).
retrieving rules such as retention and utilization goal are duly
urgent. Furthermore, we mentioned modern-day studies of
blockchain on healthcare SCM. In particular, the applicability
of blockchain to cope with consider degradation and to Design
reinforce fact transparency at the scientific trials. Several
researchers advise using blockchain to enhance the medical Recommender system
credibility of findings from scientific trials, which can be
undermined through issues including lacking facts and First, we take source datasets from the websites that
selective publication. host data about doctors. This dataset contains important
From (4), another function we have got protected is information such as consultant fees, locations where the
data-to-textual content technology systems, which are more doctors work, ratings, hospitals, etc. The dataset will be
and more used withinside the fitness domain. They can, generated using Web Scraping technology. The dataset is
for instance, be used for automation of fitness reports, then cleaned by removing unwanted outliers and missing
scientific selection support, inspiring behavioral change, values. Incorrect tuples are either removed or corrected by
making sure affected person engagement, or helping sufferers scrapping another source.
with making fitness decisions. Many sufferers are not able Some part of the cleaned data is then passed to the ML
to advantage of these established fitness records, due to the model where the model learns how to give suggestions based
fact that those files fail to speak vital records that impact the on certain parameters. The system will be using sophisticated
affected person’s expertise of those materials (e.g., the affected algorithms and will return with a ranked list of doctors. The
person’s character risks, issues, and values). Tailoring fitness remaining part of the cleaned dataset is used to test the
verbal exchange to character sufferers appears a fruitful accuracy and performance of the ML model. After testing,
solution. In addition, despite the fact that such customized the ML model is then implemented on the application.
4 Sumariya et al.

TABLE 1 | Tools and libraries used in hospital

recommender system (HRS).

Tools and libraries Descriptions

Truffle framework Allows the development of DApps on the Ethereum

network. Provides a set of tools for creating
solidity-based smart contracts. Provides a
framework for contract testing. Provides tools for
deploying Smart contracts on the blockchain. Inside
truffle, this is used to create client-side applications.
Ganache Runs a local Ethereum instance. Provides sample
accounts for use in the development and testing
processes.
Solidity Knowledge of object-oriented programming is
required to write and implement smart contracts on
the Ethereum blockchain.
Web3js JavaScript open-source library. API for interacting
with local Ethereum nodes.
Django Server for the backend of our system.
MetaMask Chrome browser extension acting as a Web 3 wallet.
Inserts Web3js libraries into the browser to allow
reading and writing on the Ethereum blockchain.
Web Front End ReactJs, which heavily relies on Web3js to connect
to blockchain nodes.

We use the local Ethereum framework, MetaMask,

Web3js, and Django to create the prototype of the
proposed framework.
We are using smart contracts for reading and writing
FIGURE 1 | Proposed design for the system.
the ranks of doctors. Currently, we are using a rank-
based system, where we assign a rank to each of
our recommended doctors, based on whether the user
During the runtime of the application, the users can
likes our recommendation or not. If the user likes the
input their parameters such as location and price range,
recommendation, the rank will increase.
and the model will be returning the preferred doctors’
Smart contract stores data in their own private part of the
list. Feedback from the users will be recorded, which
blockchain state, and every non-light node has a copy of this
will be used to determine the rank of the recommended
state, and, while usually not directly exchanging blockchain
doctors. The positive feedback will increase the rank of
states, nodes use hashes to make sure their copies of the
the doctors on the list. The negative ones will decrease
blockchain state are in sync.
the rank. The blockchain will be storing the ranks of
When a smart contract accesses its state variable, the node
the doctors.
that executes the corresponding transaction reads the value
The ranks of doctors are also affected by the latest data
of this variable from the local copy of the blockchain state.
that web crawlers will be mining. This new data will be
We searched for available datasets in research papers,
cleansed and will be used to retrain the ML model with the
Kaggle, and other resources on the Internet. But, a dataset
latest trends in the data. A periodic dataset generator will
that satisfied our requirements was not available. We found
be used that will handle all the latest dataset corrections.
a website which showed all the doctors and other details
The ML model will then be trained periodically to update its
that we require. We used web scraping to scrape the data
recommendations (Figure 1).
as per our requirements. The data were then preprocessed
to remove any null values and get the specific fields such as
doctor experience, fees, ratings, and location.
Implementation To convert the city to numeric form so that it can be used
to train our regression model, one hot encoder was used.
There are numerous tools and libraries available for DApp Ranks for each doctor were derived using the ratings and fee
development. The tools and libraries used in developing this values. A doctor with higher ratings will have a higher rank
Hospital Recommender System (HRS) are, however, briefly and for two doctors with the same ratings, the one with less
discussed in Table 1. fees will have a higher rank than the other.
10.54646/bije.2023.11 5

FIGURE 2 | Result of TFID Vectorizer for some sample parameters.

FIGURE 3 | Top 10 doctors based on the sample.

After preparing the final dataset, we trained our machine and Figure 3 shows the top 10 doctors based on the sample
learning models using the data. Experience, fees, and city parameters. After getting the top 10 doctors list from the
were the features that we used to predict the target rank model, we passed those ids to the blockchain, which then
for each target. We used TF-IDF Vectorizer and Cosine gives likes and dislikes of the top 10 doctors, accordingly on
Similarity (Unsupervised Learning) to find the most similar basis of likes, we showed the user in descending order of likes
tuples from the dataset. The top 10 most similar tuples from with the doctor with higher like at the top and the one with
the dataset matching the city are picked and sorted on the lowest at the bottom.
basis of user feedback.

Conclusion
Results and discussion
We have designed a solution for the difficulty in finding
To supply input to our models, we created a function that will doctors and hospitals for cancer patients. In our solution,
take the basic input and convert it to a form that can be given the user will have to sign in first followed by which the user
to our model. To start with, we supplied user input to the will have to provide parameters such as cost, location, and
TF-IDF Vectorizer, which converted the inputs into vectors. experience. Our solution will then use the TF-DIF Vectorizer
We then provided these vectors to the Cosine Similarity to convert the given parameters into vectors, and these
Algorithm, which showed the value of each record between vectors are then given to cosine similarity to get similar
0 and 1 based on the similarity. Figure 2 shows the result of records. It then fetches data from the blockchain and sorts
one prediction when the user entered particular parameters, similar records based on the number of likes, and the top
6 Sumariya et al.

10 doctors from the list are then recommended to the user. 3. Khezr S, Moniruzzaman M, Yassine A, Benlamri R. Blockchain
technology in healthcare: a comprehensive review and directions for
Thus, we have used blockchain with machine learning to
future research. Appl Sci. (2019) 9:1736.
improve the efficiency and authenticity of the recommender
4. Hommes S, van der Lee C, Clouth F, Vermunt J, Verbeek X,
system, which will make finding hospitals and doctors super Krahmer E. A personalized data-to-text support tool for cancer patients.
easy and reliable. Proceedings of the 12th International Conference on Natural Language
Furthermore, this can be extended to support all diseases. Generation. Tokyo: Association for Computational Linguistics (2019).
p. 443–52.
Features such as voice navigation, interpretation of medical
5. Zhang Q, Xu B, Jing H, Zheng Z. Ques-chain: an Ethereum based
reports, and medicine prediction can also act as add- e-voting system. arXiv [preprint] (2019). doi: 10.48550/arXiv.1905.05041
ons to this system. 6. McCorry P, Shahandashti SF, Hao F. A smart contract for boardroom
voting with maximum voter privacy. Proceedings of the International
Conference on Financial Cryptography and Data Security. Cham:
Springer (2017). p. 357–75.
References 7. Katuwal GJ, Pandey S, Hennessey M, Lamichhane B. Applications
of blockchain in healthcare: current landscape & challenges. arXiv
1. Chi CL, Street WN, Ward MM. Building a hospital referral expert system [preprint] (2018). doi: 10.48550/arXiv.1812.02776
with a prediction and optimization-based decision support system 8. Zhang P, White J, Schmidt DC, Lenz G. Applying software patterns
algorithm. J Biomed Inform. (2008) 41:371–86. to address interoperability in blockchain-based healthcare apps. arXiv
2. Miller A, Juels A, Shi E, Parno B, Katz J. Permacoin: Repurposing bitcoin [preprint] (2017). doi: 10.48550/arXiv.1706.03700
work for data preservation. Proceedings of the 2014 IEEE Symposium on 9. Khatoon A. A blockchain-based smart contract system for healthcare
Security and Privacy. Berkeley, CA: IEEE (2014). p. 475–90. management. Electronics. (2020) 9:94.