While collaboration is becoming increasingly critical in many industries to maximize revenue and minimize cost by taking advantage of other’s expertise, healthcare industry has hardly utilized the advantages of collaboration. This paper proposes a framework for collaborative healthcare services. The proposed framework is based on Service Oriented Architecture using web services and utilizes a medical coding system, Current Procedure Terminology (CPT), for interoperability. The framework is designed to facilitate collaborations among healthcare service providers, and is expected to reduce medical expenditure of patients and increase revenue of healthcare service providers. In order to verify the performance of the proposed framework, a simulation study is conducted. The experimental results show that the framework contributes to utilization increase and treatment lead time decrease in healthcare services.
Collaboration is a key success factor in many businesses. Manufacturers design a product together with many suppliers and outsource parts to make the product. For example, Intel, Samsung, LG and Sharp are major suppliers of parts for Apple’s flagship product—iPhone, and they collaborate even from the initial stage of a product design. The goal of collaboration is to maximize product performance and minimize costs by taking advantage of specialties of collaborating partners. Boeing’s strategy for developing the 787 Dreamliner is another good example. Boeing has partnered with fifteen companies in ten U.S. states and seven countries to create the major structural systems of the aircraft [
While collaboration has been utilized in many industries and contributed to cost reduction and performance improvement, healthcare industry has rarely utilized the advantages of collaboration. Very few literatures have addressed the collaboration issues among healthcare service providers [
Increased utilization of resources in hospitals can also help reduce medical costs.
In this paper, an information system framework that supports healthcare service collaboration among healthcare service providers is proposed and a simulation study is conducted to test the viability of the proposed framework. This paper is organized as follows: Section 2 reviews previous literature and justifies the contribution of this paper. Section 3 presents the information system architecture and key technologies for the proposed framework for healthcare service collaboration. In the section, possible liver cancer diagnosis scenarios based on the proposed framework are introduced. Section 4 discusses the result of a simulation study that supports the advantages of the proposed framework. Finally, Section 5 concludes the paper with the contribution and limitation of the proposed ideas as well as future research topics.
. Itemized MRI cost
Expense Category | Type of Expense | Operational Cost/Year | Cost per 1000 Exams | Cost per 2000 Exams |
---|---|---|---|---|
MRI Equipment | Fixed | $300,000 | $300 | $150 |
Maintenance | Fixed | $100,000 | $100 | $50 |
Space | Fixed | $50,000 | $50 | $25 |
Personnel | Fixed | $125,000 | $125 | $63 |
Overhead | Fixed | $50,000 | $50 | $25 |
Office Expenses | Fixed | $15,000 | $15 | $8 |
Marketing | Fixed | $10,000 | $10 | $5 |
Utilities | Fixed | $50,000 | $50 | $25 |
Medical Supplies | Variable | $25 | $25 | $25 |
Film | Variable | $30 | $30 | $30 |
Billing & Collect | Variable | $35 | $35 | $35 |
Total | $790 | $441 |
Collaboration in healthcare service areas, particularly at the hospital level, has not been common. Medical collaboration not only means sending medical records electronically from one hospital to another hospital, but it also means collective activities among healthcare service providers for diagnosis or treatment processes for a patient. Collaboration among hospitals can facilitate access to quality care for patients [
The utilization of expensive medical resources has been a major issue for hospitals. It is not uncommon for multiple competing MRI (Magnetic Resonance Imaging) centers to go half utilized and even far less than that [
Although collaborative healthcare services among healthcare service providers have not been common, the infrastructure for medical collaboration has been developed and implemented among medical organizations. Most hospitals utilize inter- or intra-hospital information systems (HIS), managing information on patients, appointment status, schedule of doctors and nurses, and billing. Furthermore, electronic health records (EHR) systems [
Web services have been applied to support inter-hospital communications and can facilitate collaboration among healthcare service providers and improve the quality of healthcare services [
Medical coding systems have significantly contributed to the interoperability between healthcare service providers and insurance companies. Healthcare service collaboration cannot work without interoperation among heterogeneous systems that each hospital utilizes. One of the widely used medical coding systems is CPT (Current Procedural Terminology) codes [
The CPT coding system has contributed to ensuring uniformity among medical organizations. The same CPT code used in two hospitals means the same healthcare service, such as appendectomy or endoscopy, but does not mean the same healthcare service fee. One hospital may charge a higher fee than the other hospital for the same CPT code, which means the coding system is for the identification of healthcare services, not for the decision of service fees. On the other hand, Healthcare Common Procedure Coding System (HCPCS) [
As discussed in this section, medical collaboration has not been common in current healthcare services, but underlying technologies for the collaboration has been established. Medical collaboration is expected to contribute to timely services with reduced cost for patients and revenue increase for hospitals by increased utilization. This paper proposes a framework for collaborative healthcare services. The proposed framework is based on Service Oriented Architecture using web services and utilizes the CPT medical coding system.
The proposed framework for collaborative healthcare services is based on Service Oriented Architecture (SOA) [
The conceptual diagram for the proposed collaborative healthcare service framework is illustrated in
In order for the medical collaboration broker to identify most-suitable healthcare service or a series of health
Collaborative healthcare service framework
care services, the broker needs to ascertain available services provided by the healthcare service providers that participate in the medical collaboration. The healthcare service providers share with the collaboration broker detailed information about their healthcare services. The information includes not only all the details of each healthcare service but also general hospital information, such as addresses or phone numbers. The detailed information about each healthcare service contains the CPT (Current Procedure Terminology) code for the healthcare service, availability, service lead time, doctor specializations, service fees, and etc. As shown in
Healthcare services are defined as activities for patients conducted by healthcare service providers, such as tests or operations. In order to share available resources for healthcare services seamlessly among collaborating hospitals, healthcare services need to be described precisely. A machine-readable language based on Extensible Markup Language (XML) [
In order to standardize and automate the access to the healthcare service repository, the collaboration broker should provide subscription, publishing services for healthcare service providers and query services for healthcare service users. Those services are categorized into 1) managerial services and 2) query services. Managerial services consist of web services for subscription and un-subscription of healthcare service providers. Query services are provided for health care service users.
A medical collaboration broker, healthcare service users and healthcare service providers are three main players in the proposed framework. The proposed collaboration broker system provides web services for healthcare service providers, such as publishing, updating, deleting healthcare services. In addition, the broker system provides web services for healthcare service users, such as searching for a healthcare service or generating workflow that consists of multiple healthcare services. Searching for a healthcare service can be accomplished by invoking a web service. On the other hand, the web service for workflow generation arranges multiple healthcare services in a required sequence, which is pre-specified by medical doctors. While integrating multiple healthcare services, the web service for workflow generation utilizes single web service for healthcare service search multiple times.
Healthcare service description language for bronchoscopy
UML sequence diagram for collaborative healthcare service scenarios
tionally, a patient who needs to receive multiple medical examinations would go to one hospital and undergo all four tests based on the resource availability of the hospital. In the proposed collaboration framework, however, the CPT code of the multiple medical examinations will be sent to hospitals participating in healthcare service collaboration via the proposed collaboration broker. Once the broker collects all availability information from the participating hospitals, the workflow generation module in the broker system will recommend the best healthcare service workflow based on the patient’s priority, such as time, cost, or quality.
In the proposed collaborative healthcare service framework, the healthcare service coordinator requests a query to the collaboration broker with the CPT codes of four recommended medical tests and their sequence requirements.
A simulation study is conducted to verify the performance of the proposed collaborative healthcare service framework compared to the current practice, which can be regarded as the non-collaborative healthcare service framework. This section describes the two healthcare service models, collaborative and non-collaborative service models, which are compared in this simulation study. Two performance measures are defined and used for the comparison of the two models. Finally, the simulation results and statistical analysis results are discussed in detail in this section.
Two healthcare service models, non-collaborative service model (NSM) and collaborative service model (CSM) are designed to represent the current practice and the proposed framework, respectively. The performance of the two models is compared in this simulation study. In NSM, patients have a primary healthcare service provider
Liver cancer diagnosis processes in terms of four different objectives
and receive all the treatments completed with the healthcare service provider. For example, if a patient needs a chest X-ray test and a sputum cytology test, the tests will be conducted by the service provider regardless of how early both tests are available with the service provider. On the other hand, in CSM, patients have a primary healthcare service provider, but can choose any healthcare service provider based on their needs, such as time or cost. For example, if a patient needs the two aforementioned tests at the earliest time, the patient can choose a service provider that has an earlier available schedule than other providers. During the service provider selection process, healthcare service coordinators, which are discussed in Section 3, will provide professional consulting services. These two service models are illustrated in
In the simulation study we have considered four hospitals, denoted by H1 through H4, and three different treatments, denoted by T1 through T3. Four different patient types (PT), denoted by PT1 through PT4, are defined based on the series of treatments that a patient needs to receive. The treatment sequence for Patient type 1 (PT1) is T1→T2; that of PT2 is T2→T3; that of PT3 is T1→T3; and that of PT4 is T1→T2→T3. The sequence of treatments is maintained in the simulation model by allowing the succeeding treatment to start only after completion of the preceding treatment. Inter-arrival time of patients is considered to be exponentially distributed and treatment times are constant. In order to consider various scenarios at individual hospitals, ten scenarios of different inter-arrival and treatment times are considered as shown in
Collaborative and non-collaborative healthcare service models
. Experimental design
No. | Inter-arrival Time (min) | Treatment1 (min) | Treatment 2 (min) | Treatment 3 (min) |
---|---|---|---|---|
1 | 55 | 20 | 25 | 20 |
2 | 60 | 20 | 25 | 20 |
3 | 60 | 20 | 25 | 30 |
4 | 75 | 25 | 25 | 25 |
5 | 75 | 20 | 25 | 30 |
6 | 85 | 20 | 25 | 20 |
7 | 90 | 20 | 25 | 30 |
8 | 90 | 25 | 25 | 25 |
9 | 95 | 20 | 25 | 25 |
sidered to be exponentially distributed. The detailed experimental design is given in
To compare the performance of the CSM and the NSM, two performance measures, resource utilization and treatment lead time, are used. The definition of the two performance measures are as follows.
1) Resource Utilization (RU): Percentage of resource usage in terms of time.
2) Treatment Lead Time (TLT): Total time taken to complete all the treatments for all patients including waiting time.
A statistical analysis is conducted to verify the advantages of the proposed collaborative service model (CSM) over the current non-collaborative service model (NSM). Two paired t-tests are performed to identify whether or not the collaborative model results in increased resource utilization (RU) and decreased treatment lead time (TLT) against the non-collaborative model. The null and alternative hypotheses for treatment lead time are H0:
. Experimental results
No. | Resource Utilization (%) | Treatment Lead Time (minute) | ||||
---|---|---|---|---|---|---|
Non-collaborative | Collaborative | Difference (%) | Non-collaborative | Collaborative | Difference (%) | |
1 | 87.0 | 88.7 | 1.9 | 200 | 132 | 33.8 |
2 | 81.2 | 83.0 | 2.1 | 145 | 100 | 30.8 |
3 | 91.3 | 92.9 | 1.7 | 309 | 206 | 33.4 |
4 | 74.9 | 75.5 | 0.8 | 139 | 101 | 27.4 |
5 | 79.8 | 80.5 | 0.8 | 174 | 117 | 32.6 |
6 | 58.1 | 58.2 | 0.2 | 143 | 70 | 51.4 |
7 | 63.3 | 63.4 | 0.2 | 104 | 82 | 20.9 |
8 | 63.4 | 63.5 | 0.1 | 104 | 84 | 19.5 |
9 | 56.2 | 56.3 | 0.1 | 85 | 72 | 15.2 |
Average | 0.9 | Average | 29.4 |
(TLT) and more resource utilization (RU) than does the NSM.
On average, 29.4% reduction in treatment lead time (TLT) is expected by utilizing the collaborative service model (CSM) over the non-collaborative service model (NSM). Particularly, Scenarios 1 through 6 have much higher reduction in TLT than the other scenarios. This finding shows that hospitals with higher utilization in NSM can expect higher reduction in TLT than hospitals with lower utilization. On the other hand, in terms of resource utilization on average 0.9% increase in resource utilization is expected. In particular, hospitals with higher resource utilization, shown in Scenarios 1 through 5, have relatively higher increase in resource utilization than hospitals with lower resource utilization.
This paper proposes a framework for collaborative healthcare services in order to facilitate collaborative services among hospitals to provide reduced treatment lead time for patients and increased utilization of medical resources of hospitals. The service scenarios are discussed in detail, which take into account four different considerations of patients to show how healthcare services can be provided based on the proposed framework. In order to verify the advantages of the proposed framework, a simulation study is conducted. The current practice of the non-collaborative service model (NSM) and the proposed framework of the collaborative service model (CSM) are compared using paired t-test. The simulation experiment results show that the proposed framework can contribute not only to the increase of hospital resource utilization, but also to the reduction of treatment lead time.
The Current Procedure Terminology (CPT) code plays a key role in the framework. The framework provides various services, such as searching medical services, integrating them and generating workflows, using the coding system. The unique identification capability of the CPT code enables accurate search for services provided by healthcare service providers. Although the proposed architecture adopts the Current Procedure Terminology (CPT) code to identify specific healthcare services, the architecture of the proposed framework is not dependent on the CPT coding system and can easily adopt alternative coding systems. The coding system has kept improving to reflect the practical needs. As the coding system improves, the proposed framework is expected to provide more accurate services.
As a future research topic, a mathematical model should be designed to generate the optimal treatment plan according to the objective of the service users. The mathematical model should be able to incorporate multiple objectives because healthcare-related decisions are normally made considering multiple aspects. These mathematical models should be integrated with the real-time availability information provided by healthcare service providers via the proposed healthcare collaboration broker. Considering the complexity of the treatment plan generation, a heuristic algorithm should also be developed to provide real-time collaborative healthcare services. If the treatment plan is urgently required, the heuristic algorithm can be utilized. Otherwise, the mathematical model can be used. The mathematical model can be also used as a benchmark tool for the performance of the heuristics algorithm.