IN
DEGREE PROJECT MEDICAL ENGINEERING, SECOND CYCLE, 30 CREDITS
STOCKHOLM SWEDEN 2020,
A Review of Success and Failure Factors of using Patient-Generated Health Data for Chronically Ill
Patients
LEJLA DELILOVIC
KTH ROYAL INSTITUTE OF TECHNOLOGY
SCHOOL OF ENGINEERING SCIENCES IN CHEMISTRY, BIOTECHNOLOGY AND HEALTH
This master thesis project was performed in collaboration with Cambio Healthcare Systems
A Review of Success and Failure Factors of using Patient-Generated Health Data for Chronically Ill Patients En överblick av inflytelserika faktorer av patient-genererad hälsodata
för kroniskt sjuka patienter LEJLA DELILOVIC
Master of Science Thesis in Medical Engineering Advanced level (second cycle), 30 credits Supervisor at Cambio: Terese Karlsson
Supervisor at KTH: Maksims Kornevs Examiner: Sebastiaan Meijer TRITA-CBH-GRU-2020:035.
KTH Royal Institute of Technology School of Engineering Sciences in Chemistry, Biotechnology and Health (CBH) SE-141 52 Huddinge, Sweden http://www.kth.se
Abstract
Our population is becoming older and with that, the development of chronic dis- eases is also expected to increase. A chronic illness is a long-term illness which lasts throughout a lifetime, or at least for a very long time. A large part of healthcare resources is already devoted to treating chronically ill patients. These patients are often dependent of both care and medication to maintain a meaningful life. To gain a holistic view of these patients health condition by one/two appointments with physicians yearly is not sufficient in order to conclude a certain health-state.
The course of disease for these patients changes daily and require follow-up on dis- ease progression continuously to adapt an appropriate treatment plan. Collecting patient-generated health data (PGHD) facilitates in the process of retrieving more evidence for better assessment of the disease development. While there is obvious importance and benefit of using of PGHD, this data is not commonly used in health- care. Further investigation is needed to understand how PGHD can be more useful.
This pilot study provides knowledge of the success and failure factors of using PGHD for mainly chronically ill patients, but can be applied to other patient-groups as well.
The aim of this thesis work was to collect information about what suppliers, gov- ernmental organizations and healthcare professionals require for using PGHD in healthcare setting in a greater extent in the future. Methods used to gather in- formation were participatory interviews in combination with qualitative interview questions. Pattern recognition has been created through a thematic analysis and cluster mapping. The data collection resulted in four areas of improvement; patient behaviour, healthcare organization, digitized health data and equipment.
The result shows overall a positive attitude towards the concept of PGHD by all sec- tors asked in this project. Stakeholders agree on that PGHD can generate positive outcomes for chronically ill patients. The belief of improving workflow in healthcare with PGHD was also positive. The valuable possibilities generated with PGHD are tailored care flows, improved evaluation of disease status and enhanced quality of care and well-being among others. Additionally, several ongoing projects are taking place, which demonstrate great interest in the area. However, before PGHD can be prescribed by healthcare, studies have to be performed including development of national guidelines for reporting PGHD, building a secure infrastructure and in- troducing new work routines. Future work will be applying AI-analysis of reported PGHD to facilitate the work of caregivers and development of secure storing solu- tions for instance with block-chain technology.
Keywords
Patient Generated Health Data, PGHD, Self-Care, eHealth, Chronic Diseases
Sammanfattning
V˚ar befolkning blir ¨aldre och med det f¨orv¨antas utvecklingen av kroniska sjukdo- mar ocks˚a att ¨oka. En kronisk sjukdom kallas ocks˚a f¨or en l˚angvarig sjukdom som varar under hela livet eller ˚atminstonde under mycket l˚ang tid. En stor del av v˚ardresurserna ¨agnas redan ˚at att behandla kroniskt sjuka patienter. Dessa pa- tienter ¨ar ofta beroende av b˚ade v˚ard och medicinering f¨or att uppr¨atth˚alla ett meningsfullt liv. F¨or att f˚a en helhetssyn p˚a dessa patienters h¨alsotillst˚and kr¨avs mer ¨an en eller tv˚a m¨oten med sjukv˚arden ˚arligen. Ett f˚atal m¨oten per ˚ar ¨ar inte tillr¨ackligt f¨or att konstatera en r¨attvis bild av h¨alsotillst˚andet. Sjukdomsf¨orloppet f¨or dessa patienter f¨or¨andras dagligen och kr¨aver fortl¨opande uppf¨oljning av sjuk- domsutvecklingen f¨or att f¨orst˚a och anpassa en l¨amplig behandlingsplan. Insamling av patientgenererad h¨alsodata (PGHD) underl¨attar och hj¨alper till i denna process.
Det finns uppenbarliga f¨ordelar med PGHD, men datan anv¨ands inte vanligtvis i sjukv˚arden. Ytterligare forskning beh¨ovs f¨or att f¨orst˚a hur PGHD kan vara mer anv¨andbart.
Denna pilotstudie ger kunskap om inflytelserika faktorer f¨or att anv¨anda PGHD f¨or huvudsakligen kroniskt sjuka patienter, men kan ocks˚a till¨ampas p˚a andra pa- tientgrupper. Syftet med detta avhandlingsarbete var att samla in information om vad leverant¨orer, statliga organisationer och v˚ardpersonal kr¨aver f¨or att anv¨anda PGHD i v˚ardmilj¨o i framtiden. Metoder som anv¨ants f¨or att samla in informa- tion var deltagande intervjuer i kombination med kvalitativa intervjufr˚agor. Re- sultatet har genomg˚att en klusterkartl¨aggning och tematisk analys f¨or att skapa m¨onsterigenk¨anning. Datainsamlingen resulterade i fyra omr˚aden; patientbeteende, v˚ardorganisation, digitaliserad h¨alsodata och utrustning.
Resultatet visar en positiv inst¨allning till PGHD enligt alla deltagare i studien.
Deltagarna ¨ar ¨overens om att PGHD kan generera positiva resultat f¨or kroniskt sjuka patienter. Tron att f¨orb¨attra arbetsfl¨odet inom h¨also-och sjukv˚ard med PGHD var ocks˚a positiv. De v¨ardefulla m¨ojligheterna som genereras med PGHD ¨ar bl.a.
skr¨addarsydda v˚ardfl¨oden, f¨orb¨attrad utv¨ardering av sjukdomstatus och f¨orb¨attrad kvalitet p˚a v˚ard och v¨alm˚aende bland andra. Dessutom p˚ag˚ar flera p˚ag˚aende pro- jekt som visar stort intresse f¨or omr˚adet. Innan PGHD kan b¨orja f¨orskrivas av h¨alsov˚arden m˚aste det utf¨oras fler studier som inkluderar framtagning av nationella riktlinjer f¨or rapportering av PGHD, byggandet av en s¨aker infrastruktur och in- troduktion av nya arbetsrutiner. Framtida arbete kvarst˚ar d¨ar till¨ampning av AI- analys modeller p˚a rapporterad PGHD samt utveckling av s¨akra lagringsl¨osningar, f¨orslagsvis med blockkedjeteknik, b¨or vidareutvecklas.
Nyckelord: Patient-genererad h¨alsodata, PGHD, Egenv˚ard, e-h¨alsa, Kroniker
Acknowledgements
This master thesis project has been performed at the School of Engineering Sciences in Chemistry, Biotechnology and Health (CBH) at KTH Royal Institute of Tech- nology in collaboration with Cambio Healthcare Systems. The thesis is the most comprehensive and interesting work I have done during my studies at the Medical Engineering program at KTH.
I would like to thank all interview participants who found time in their busy sched- ules to take part in my interviews. Your participation has meant everything to the results of this project. Our discussions have been tremendously valuable and I have learnt so much during this time.
I would especially like to thank my supervisor, Terese Karlsson at Cambio, for proofreading, helping me to get in contact with relevant people and supporting me with great ideas throughout the project. Thank you for all the time and support I have received from you. It has truly meant a lot to me.
I would also like to thank my KTH supervisor, Maksims Kornevs, for guiding and helping me throughout the work. I am also thankful for my examiner, Sebastiaan Meijer, for your feedback on my work.
I would like to thank my friends for the support I have received. Especially, I would like to thank Rebecka Hansson for proofreading and providing valuable feed- back on my report.
Finally, I would like to thank my mom, dad and my beloved Olof. Thank you for always believing in me and for the endless support I receive from you every day.
It means more than you can ever imagine.
Sincerely,
Lejla Delilovic
KTH Royal Institute of Technology
Contents
1 Introduction 1
1.1 Aim . . . . 2
1.2 Research Questions . . . . 2
1.3 Summary . . . . 2
2 Background 3 2.1 Digital Health . . . . 3
2.1.1 Remote Patient Monitoring . . . . 4
2.1.2 Mobile Health Technology . . . . 4
2.2 Digitalization in Swedish Healthcare . . . . 4
2.3 Patient-Generated Health Data . . . . 6
2.4 Influencing Factors of PGHD . . . . 7
2.5 Synopsis . . . . 8
3 Methodology 9 3.1 Design of Data Collection . . . . 9
3.1.1 Literature Study . . . 10
3.1.2 Participatory Method . . . 10
3.1.3 Interview Questions . . . 11
3.2 Data Analysis . . . 11
3.2.1 Thematic Analysis . . . 12
3.2.2 Validation . . . 13
4 Results 14 4.1 Data Collection . . . 14
4.1.1 Literature Findings . . . 14
4.1.2 Distribution of Respondents . . . 15
4.1.3 Participatory Method . . . 16
4.1.4 Quantitative Questions . . . 19
4.2 Data Analysis . . . 21
4.2.1 Thematic Analysis . . . 21
4.2.2 In-Depth Analysis of Factors . . . 22
4.2.3 Qualitative Questions . . . 23
5 Discussion and Conclusion 31 5.1 Discussion of Methodology . . . 31
5.2 Discussions of Results . . . 32
5.2.1 Comparison of Literature and Findings . . . 32
5.2.2 Knowledge and Interest . . . 34
5.2.3 Current Work . . . 35
5.2.4 Value of PGHD . . . 36
5.2.5 Healthcare Flow . . . 38
5.2.6 Barriers of PGHD . . . 39
5.3 Future Work . . . 40
5.4 Conclusion . . . 41
References 43 A Appendix A i A.1 Interview Questions . . . . i
B Appendix B iii B.1 Application Areas . . . . iii
B.2 Standardization and Interoperability . . . . iv
List of Figures
2.1 Chart of the Swedish Health Eco-System. . . . 5
2.2 Relationship between personal and organizational domain. . . . 7
3.1 Chart of the Methodology . . . . 9
3.2 Example of a cluster relationship between different themes (circles) and their factors (rectangles) [35]. . . 13
4.1 Three notes created by a candidate during an interview. . . . 16
4.2 Graphical distribution of general knowledge in PGHD. . . . 19
4.3 Graphical distribution of general interest in PGHD. . . 20
4.4 Likert scale with four alternatives for each question. . . 20
4.5 Graphical distribution of general work with PGHD. . . 21
4.6 A thematic analysis mapping of the factors from the template. . . 22
5.1 Top-bottom approach of the thesis project. . . 32
List of Tables
3.1 Compilation of keywords respective quantity of hits. . . 10
3.2 Example of structuring codes, factors and domains. . . 12
4.1 Factors from literature study. . . . 14
4.2 Distribution of interview candidates. . . . 15
4.3 New factors created by interview participants. . . 17
4.4 Additional factors created by interview participants. . . 18
Acronyms
EES Electronic Expertise Support EHR Electronic Health Record EMR Electronic Medical Record HIS Healthcare Information System HTO Human Technology Organization
ICT Information and Communication Technology IoT Internet of Things
IVO Inspektionen f¨or v˚ard och omsorg (Inspectorate for Health and Care) mHealth Mobile Health Technology
PGHD Patient-Generated Health Data PHR Patient Health Record
RPM Remote Patient Monitoring
SKR Sveriges kommuner och regioner (Sweden’s Municipalities and County Coun- cils)
Nomenclature
Clinical Informatics Enabling the use of information in healthcare by and for clinicians
Health Informatics Methods and techniques for collection, management and eval- uation of medical information
Medical Informatics Intersection of information science, computer science and healthcare
Telemedicine Services/tools allowing long-distance contact between patient and careprovider
Chapter 1 Introduction
Swedish healthcare is considered to be one of the most developed healthcare systems in the world [1]. The law states that each Swedish citizens have right to equal care according to the needs required [2]. The chance of surviving difficult operations and preventing premature death is great in Sweden [1]. This has partly to do with the fact that Swedish healthcare have access to excellent healthcare equipment, a wide selection of medicine and high life expectancy [3]. However, healthcare is facing ma- jor challenges in the future due to ageing and growing population, increased obesity and lack of physical activity [4]. All these factors contribute to an increased num- ber of patients receiving difficult illnesses which can result in demanding treatment plans and higher costs for the healthcare system. Moreover, increase in population diseases leads to growing healthcare queues and contributes to making healthcare feel inaccessible. The trust in availability of healthcare and reasonable waiting time for medical care has fallen among the population [5]. At the same time, accessibility requirements have increased. People are seeking healthcare for less serious ailments, which also contributes to an increase in waiting times. In order to improve this situation healthcare might benefit of implementing digitalization tools.
Increasing engagement in personal health and interest in monitoring technology with an estimated number of 70 million wearable devices sold in 2018, demonstrates a trend in health tracking and adoption of mobile technology (mHealth) among people worldwide [6]. The next step is to digitize Swedish healthcare in order to utilize resources along the entire healthcare chain. One way to achieve this can be by using patient-generated health data (PGHD).
PGHD is health-related data collected with mHealth, by patients or caregivers, to address a health issue, for example a chronic disease. Almost half the population in Sweden suffer from a chronic disease [7]. The most common diseases in Sweden include diabetes, cardiovascular diseases and cancer [8]. A large proportion of health care funding is budgeted to treating chronically ill patients, all of 85% of total healthcare costs [9]. These high costs need to be overlooked and reduced with help of digitalization with PGHD. PGHD is already introduced to the market and now at the growth level in its adoption-curve, is expected to mature until 2024 [10].
Studies have shown that PGHD is especially applicable to patients living with chronic diseases for several reasons. This patient group have limited access to health- care despite their constant state of illness. The monitoring during a counseling ses- sion contributes to only a fraction of these patient’s health status. By using PGHD you can extend this view outside of clinical setting, enabling patient-centered care[6].
1.1 Aim
This project aims to list and investigate the underlying explanations as to why and which success and failure factors there are of using PGHD for chronically ill patients.
The goal of this study is to present a compilation of important factors worth taking into consideration when developing a framework for PGHD. The study concerns the use of PGHD for chronically ill patients but can be applied on other patient- groups as well. The expectation is to establish a basis to help in future realization of technological infrastructure created by stakeholders within health informatics.
1.2 Research Questions
The project aims to answer the following research questions:
• What is the general knowledge and interest of PGHD among stakeholders within health informatics?
• What potential value can PGHD create to improve workflow in healthcare?
• What main factors need to be considered in the implementation of PGHD?
1.3 Summary
In summary, studies have demonstrated many benefits of using PGHD such as sup- port in clinical decision-making, increased quality of life and facilitated coordina- tion of care [11], [12], [13]. Although previous studies have been done there are gaps in the area of PGHD that have yet not been thoroughly investigated. Despite an increased interest in connected care and telemedicine, there is limited empiri- cal research which has actually been accomplished in the topic, especially from the perspective of healthcare profits and technical implementation. In order to gain a holistic view of the potential that PGHD can contribute to, there is a need to fur- ther research to understand full potential of implementing PGHD. This project will therefor illuminate the success and failure factors to unlock the potential for PGHD reaching healthcare.
Chapter 2 Background
This chapter intends to put the study in a broader context, give the subject depth and covers what has previously been done in the subject area. The chapter includes rele- vant scientific background of important components in the subject area such as digi- talization and eHealth, Swedish healthcare and previous studies on patient-generated health data.
2.1 Digital Health
The digitization era began during the end of 19th century, introducing the ”World Wide Web” and the first digital computer [14]. The healthcare organization em- braced the digital health landscape as well as many other industries underwent digital infrastructure changes at that time. The healthcare shifted from a fairly stable to a more dynamic organization. Today, there are currently several actors working from different directions with digital health, also called eHealth.
eHealth is a broad term which encompasses many areas such as health infor- matics, medical informatics and clinical informatics. The concept of eHealth is a common name for the use of information and communication technology (ICT) in healthcare environment including the development and implementation in the sys- tem, as well as the design and evaluation. The concept of eHealth can be divided into three main areas: technical and social infrastructure, repositories and applications [15].
Technical and social infrastructure refers to the data exchange between actors in the healthcare structure for example between care at home, hospitals and healthcare centers, regardless of public or private sector. Standardization, interoperability and semantics fall under this area as well as regulatory and ethical aspects [15].
eHealth repositories are for example Electronic Health Records (EHR) and Per- sonal Health Records (PHR) where health data is stored digitally in secure databases.
These resources are valuable in the sense of allowing easily accessible health infor- mation for authorized personnel to facilitate in the management of care-giving [15].
eHealth applications are services and tools which purpose is to support healthcare flow and to gather patient readings. The building blocks are telemedicine and various ICT applications [15].
Vision eHealth 2025 is a goal in Swedish healthcare which expresses that Sweden should work towards being at the forefront of using digitalization opportunities in healthcare by year 2025 [16]. The eHealth Authority is a department working
towards accomplishing the Vision e-Health goal 2025. Important projects driven by the eHealth Authority are the National Drug List and Electronical Expertise Support (EES). A main part of their work is also to quality-assure infrastructure connected with healthcare providers and residents [17]. Assuring this infrastructure ensures a patient-safe healthcare in home environment as well. Remote patient monitoring is therefore a fast growing area since the possibilities of safe patient monitoring are increasing with eHealth developments.
2.1.1 Remote Patient Monitoring
Remote patient monitoring (RPM) uses digital technologies to monitor patients with chronic diseases in a non-traditional way outside of healthcare setting. Healthcare provides the necessary equipment to the patients home in order to monitor and collect health-data, which is known as home-monitoring. The purpose of RPM is to improve decision-making and save time for both patients and healthcare [18]. A prerequisite for efficient RPM is well-developed monitoring, trust in remote technol- ogy and patient involvement. The outcome is hopefully that it can contribute to a reduced number of emergency hospital admissions by detecting deterioration faster and deploy care resources. One of the most common way to monitor elderly people in Sweden at home is currently through home monitoring.
2.1.2 Mobile Health Technology
Mobile health technology (mHealth) is an entity in eHealth and commonly used in RPM. The area includes mobile components such as smart-phones, tablets, wrist- bands and other wireless devices measuring physical parameters. The collective name for these technologies is Internet of Things (IoT). These devices support the gathering of medical information data to improve patient care and enhance quality of diagnosis determination among others [19]. mHealth is fundamental in the digi- tization process since it provides the ability to collect health information regardless of time and place. It is a powerful tool in the process of collecting data in a non- traditional way and it allows a shift in the development of healthcare moving towards an available, patient-centered and digitize community [11]. Data collected through mHealth is called PGHD. Several studies have been conducted using mHealth in hospital environment showing promising potential. However, the continuous use of mHealth applications in clinical setting is still almost non-existing [20], [21].
2.2 Digitalization in Swedish Healthcare
The Swedish healthcare system is structured into three levels: governmental, re- gional and municipal. All levels are controlled by democratically elected Swedish politicians. The governmental level is based in the Social Ministry while the regional level as well as the municipal level, consists of politicians chosen by the county’s res- idents [22].
The Social Ministry consist of, among others, these supporting units: The Public Health Authority, e-Health Authority, Health and Medical Responsibility Commit- tee and Inspectorate for Health and Care (IVO). These authorities are responsible
for working towards the goals stated by the Parliament and Government regarding health policy in Sweden [22].
Sweden is divided into 21 regions, each with a responsibility to organize health- care in their specific region. Regions are supported by Sweden’s Municipalities and County Councils (SKL) [22]. There are 290 municipalities in Sweden [23]. Each municipal is in charge of their local elderly care, patients with physical or mental disabilities, post-operative care and school healthcare [22]. SKL is the organization which brings together the 290 municipalities and 21 county councils. Their mis- sion is to function as network for sharing knowledge in the area and supporting the municipals and regions [24].
A simplified description of the Swedish health eco-system is found in Figure 2.1 below. This eco-system is dependent on a collaboration of shared information exchange to enable cooperation for delivery of efficient and accessible care, regardless of location in Sweden. One of the most important building blocks to enable this exchange is the healthcare information system (HIS). A more detailed explanation of the HIS construction and common standards used in the system, is found in Appendix B. The HIS-systems used in the Swedish healthcare are Melior, COSMIC, Take Care, NCS Cross and VAS [25]. It is an open market in motion where there is currently a new modern HIS, Milennium, being introduced by the company Cerner [26].
Digitalization in Sweden is mainly controlled by driven projects at the Social Ministry, eHealth Authority and SKL. It is of high importance to understand these levels of the health-eco system as it is a prerequisite for understanding how PGHD can develop to be used to a greater extent in society.
Figure 2.1: Chart of the Swedish Health Eco-System.
The Healthcare Chain
The healthcare organization is a large, complex and quite ineffective system al- though it is a highly necessary part of our society. The organization consists of healthcare professionals, suppliers and governmental organizations, all working to achieve efficient and satisfactory patient care. This requires cross-border coopera- tion between the various actors in the healthcare chain when developing and working with digitalization tools [27].
Digitalization for self-care management is one of the 14 key areas which have potential to bring great benefits in the chain for caregivers and patients. Self-care strives to expose the entire disease state of the patient. This approach shifts the patient from passive to engaged and demands the healthcare to focus on needs rather than resources. With the help of digital health technology and applications, self-care is both facilitating and safe. The possibility to create individual care programs is provided, reducing potential need for advance care at hospitals and increases quality of life for the patient. Moreover, self-care development could benefit monitoring of medication and preventive public health work [28].
2.3 Patient-Generated Health Data
Patient-generated health data (PGHD) is defined as ”health-related data created, recorded, or gathered by or from patients (or family members or other caregivers) to help address a health concern” [29]. Examples of PGHD are bio-metric data (blood sugar, weight and heart rate) and patient-centered data collecting personal information of pain symptoms, quality of life and physical activity of the patient [6]. It is important to separate PGHD from data generated in clinical environments.
The main difference is the ownership of the data. PGHD is owned by the patient generating the data, unlike data created in health care sites and stored in the patients EHR, which is owned by the healthcare. This is stated by a law, the Patient Data Act, which can be found in the Swedish regulation [30].
The Patient Data Act states that healthcare providers are required to keep med- ical records for each patient while the patient have the right to access their infor- mation written in their personal record. The access is not required to be via direct access, but can be. If the care-provider enables access directly to the EHR there is a need to ensure safe identification and authorization. The health data has internal confidentiality, meaning, only care-givers involved in the care and treatment of the patient should have access to the information. Patients have the right to block their personal record which means that other care providers do not have direct access to them. Some governmental organizations such as IVO, the Social Insurance Of- fice (F¨ors¨akringskassan) and the Social Ministry are always entitled to read medical records in cases that require it without consent [30]. Patients do not have the right to write in their medical record. However, there is a function in the EHR (e-Journalen in Swedish) called ”Patient’s Notes”. The function allows a patient to write own notes, for example measurements or symptoms. On the contrary, care-givers have no obligation to read it. Thus, this function is only availiable in one region in Sweden for now, with shared views on the function among caregivers [31].
There are two essentially different ways to collect PGHD. Quality assured equip- ment can be prescribed by the healthcare in order to measure relevant parameter
for specific patients or data can be collected through health tracking applications, for example FitBit and Garmin watches. PGHD is commonly collected through to- day’s mobile technology, mHealth. However, PGHD is not an entirely new concept.
Patients have throughout many decades recorded and shared their health data with healthcare in sessions. Moreover, as the growing market regarding affordable track- ing devices is exploding as well as the growing interest in personal health records (PHR) there is no doubt that we detect a trend pointing at the use of PGHD more frequently in the future [11].
Figure 2.2: Relationship between personal and organizational domain.
2.4 Influencing Factors of PGHD
The literature addresses several factors affecting the capacity of PGHD to be used in clinical setting. The use of PGHD is based on assumptions about it improving healthcare overall. A successful factor, in this report, will be considered as an element bringing many positive events while failure factors get stuck in barriers of different kinds.
The most mentioned factor in literature of using PGHD is the ”patient engage- ment” in their health and willingness to monitor and report their health data. Suc- cessful outcomes of using PGHD includes caregivers retrieving a ”comprehensive picture of the patients disease, improving quality of care, understanding patient ex- perience, increasing productivity of clinical trials and predicting treatment toxicities”.
Moreover, PGHD-use ”improves clinical outcomes, enables shared decision making, strengthens patient-provider communication and enhances care coordination through continuously updated patient information” [6], [11], [13], [32].
The less successful factors creating barriers for PGHD to reach the healthcare are ”technical infrastructure problems such as collecting and storing this data, in- tegrating it into HIS-systems and low level of interoperability of devices”. Further, we have ”security and authorization problems” as well as ”standardization issues.”
Another problem which occurs is whether PGHD will be of ”high enough quality for
interpretation, workload concerns regarding reviewing this data and financial impact of PGHD” [6], [11], [13], [32].
2.5 Synopsis
According to previous studies there are room for improvements within the area of PGHD. PGHD interacts with multiple parties: it is created by patients, interpreted by healthcare professionals and stored by a third party. This process is complicated and depending on several interacting components to work together according to a human-technology-organization perspective (HTO) as for the majority of systems used in healthcare environment. The literature presents factors of PGHD, both successful and unsuccessful, which are important to look over when investigating the area. This project will investigate what more factors exist as well as confirm the factors addressed by literature. Further investigation needs to be done in order to see what the underlying causes as to why PGHD is not used to its full potential are and how they can be processed.
Chapter 3
Methodology
This section presents the applied methodology used in this study to design a compi- lation of important areas of PGHD. The methodology consists of a literature study, design of interviews, data collection methods, data analysis methods and validation.
Figure 3.1: Chart of the Methodology
3.1 Design of Data Collection
The data collection in this study is based on the literature study and a partici- patory method in combination with interview questions to be able to answer the research questions stated in Chapter 1. Factors found frequently in the literature search were stated as important factors to look over. A number of experts, suppliers and healthcare professionals from different companies, organizations and authori- ties where asked to participate in individual interviews. The goal was to find 4-5 interview candidates from each group, selected based on their work experience and knowledge in the project topic. The interview candidates were selected from Cambio Healthcare Systems, Philips and GE Healthcare (suppliers), Karolinska Sjukhuset
Huddinge and Solna (healthcare organizations) and The eHealth Authority, SKR, Region Stockholm and the Social Board (governmental agencies) to receive a mix- ture of inputs from different backgrounds to achieve diversity of experience and knowledge in PGHD. Additionally, a representative from Gothenburg University was added to the interviews after recommendation from the Social Board.
3.1.1 Literature Study
An extensive literature study was executed in the beginning of the project to re- trieve knowledge in the area and to decide the most appropriate methodology. The developed methodology can be found in Figure 3.1 above. The selected databases for literature review were PubMed and KTH Primo, to find recent articles and up- to-date research. Additionally, information was collected from governmental reports and case studies. The initial search strategy was to use a variation of key words to identify articles of interest. The key words used included: patient-generated health data, PGHD, health informatics, factors, mHealth, eHealth, healthcare information systems, digitalization, challenges, self-care among several others. Research articles used in this study were restricted to be no longer than 20 years old. A summary of the combination of keywords and the number of hits for each combination is pre- sented in table 3.1 below. Moreover, the research question was developed from the identified gaps in the literature.
Keyword(s) PubMed Quantity
Digital Health 20 830
eHealth 31 712
Patient generated health data and Factors 1371
Keyword(s) KTH Primo Quantity Patient generated health data 455 418
PGHD 188
Patient generated health data challenges 195
Table 3.1: Compilation of keywords respective quantity of hits.
3.1.2 Participatory Method
The research process when applying participatory methods requires to conduct data during interviews with an involving approach. One way to achieve this can be by in- troducing a tool to the interview candidate which can be used during the interview to produce data in addition to answering the interview questions. Participatory research methods have shown benefits such as higher level of understanding the researchers questions and increase empowerment as well as confidence of the inter- viewee in their answers. Thus, this approach opens up for discussion between an interviewer and the interviewee [33]. The purpose with this method was to uncover different opinions and let participants use their knowledge to create something un- expected and new. Hopefully, this method will help in the assessment of detecting
correlation between the predetermined factors in the literature study in comparison to the factors created by interviewee’s. Furthermore, the selected participants in this study were required to be reasonably knowledgeable in the subject and specifically chosen based on their skills. The participatory method was divided into two phases:
Phase 1 and Phase 2 which are explained below.
Phase 1 (”Creating new factors”): The first phase began with the participants being instructed to write a few relevant factors on post-it notes which they thought were important to analyze regarding PGHD. If the participant did not understand the task, one or two prepared post-it notes with factors already written on them were presented to inspire the participant. Further, the participant presented and motivated why they had chosen the specific factors including picking the three most important ones. The complementary interview questions were asked the remaining interview time. The purpose in the first phase was to create new factors and discuss how they were relevant within the area.
Phase 2 (”Validating factors”): The last three interview candidates were asked to validate the factors generated in phase 1. The purpose here were to understand if there was shared or divided opinions regarding the factors and their connections to help in understanding the context. To avoid manipulation there were also a opportunity for participants to individually write down own relevant factors. The complementary interview questions were asked before introducing and validating the factors.
3.1.3 Interview Questions
The purpose of the interview questions was to help answer the research questions stated in section 1.2. The interview type was chosen as semi-structured, meaning some interview questions were formulated beforehand and asked to all participants while the post-it notes contributed to an open dialogue. This technique was chosen to achieve time-efficiency and a certain degree of accuracy in the assessment. The design of the questions were mostly open-ended questions but also some close-ended depending on suitable outcome. Open-ended questions were asked when desire was to receive motivations and see behaviors which were not expected. To avoid bias the close-ended questions were limited to only be asked when enough qualitative research had been collected. This design resulted in both quantitative and qualitative data.
The interview questions were founded on the underlying factors stated in section 2.4. The interview questions are presented in Appendix A.
3.2 Data Analysis
The data analysis in this study was reviewed simultaneously with the data collection.
The interviews were partly verbatim transcribed and analyzed. Most information was written down on paper during the interviews, using the transcript as support.
The two methods used in the analysis of data were thematic analysis and cluster mapping. The first step in the analysis was to detect and mark relevant information from the interview transcript. Further, most important factors were collected and divided into categories to be able to interpret and discover interesting causal and
correlating relationships. Transcript, category subdivision and highlighting were done in a simple Word document. Identification of themes were based on character- istics in the HTO-concept. The concept is used in processes between humans (H), technology (T) and organization (O). The idea is to apply this concept in order to detect relationship between acitivites, processes and performances in a holistic view of implementing PGHD.
3.2.1 Thematic Analysis
One of the most common analysis strategies in qualitative research design is the- matic analysis. The method is used for identifying, analyzing and reporting patterns within an amount of data. The method process includes observing important in- formation which is then divided into several themes depending on relevance in each theme-category. To perform the thematic analysis in a structured method the six steps introduced by Braun et al. was used:
1. ”Familiarizing the data.
2. Generating initial codes.
3. Searching for themes.
4. Reviewing themes.
5. Defining and naming themes.
6. Producing the report” [34].
Template Analysis
Code Factor Domain
Quality of Life Motivational pathway Human
Integration into HIS Barrier Technical
Workload concerns Organizational
X Y Z
Table 3.2: Example of structuring codes, factors and domains.
Next step was identifying factors from the codes received from the interviews.
Finding factors, helped in the search for themes, was facilitated by writing all initial codes, taken from the literature study, on post-its and structuring them into do- mains: human, technology and organization (HTO). Thereafter, a digital mind-map was created containing all factors and themes. This approach gave insight in the underlying construction on how codes could be translated into factors which were used to define the themes with the help of a template analysis according to table 3.2.
The initial themes were associated with the domains in the template analysis.
However, new codes, factors and domains were generated during the interview ses- sions. All of them are presented in the result section. The possibility of factors
partially overlapping different themes and domains had to be considered. Therefor, the cluster template in Figure 3.2 below was used as inspiration when producing the cluster map [35].
Figure 3.2: Example of a cluster relationship between different themes (circles) and their factors (rectangles) [35].
3.2.2 Validation
Validation of the interviews was carried out by emailing a summary of what had been discussed during the interviews to all participants to be confirmed. The answers to the questions general interest, knowledge and work in PGHD were unclear in the interviews. They were emailed again with response options created from a Likert scale to all participants to ensure clear and comparable answers. The results can be found in Figures 4.2, 4.3 and 4.4.
Chapter 4 Results
4.1 Data Collection
This section presents the results of the thesis, both the primary and secondary data.
The primary data includes the findings from already published material found in the literature search. The secondary data includes the result generated from the semi-structured-interviews including phase I and II. In summary, the data collection consist of the result from the participatory method and interview questions. The data is presented in several tables and sections below.
4.1.1 Literature Findings
The main result from the literature study are the factors identified as important to look over in the process of finding value from PGHD. The factors are presented randomly in table 4.1 below.
Factors from Literature Study Patient engage-
ment
Comprehensive picture of dis- ease condition
Quality of care
Understanding patient experi- ence
Financial impact Productivity of clinical trials Predicting treat-
ment toxicities
Improve clinical outcome
Strengthen patient-provider communication Integration into
HIS
Enable shared decision making
Enhance care co- ordination Technical infras-
tructure
Collecting and storing data
Interoperability of devices Workload con-
cerns
Interpretation of data
Security and au- thorization Standardization
of data
Time-efficiency Culture manage- ment
Table 4.1: Factors from literature study.
4.1.2 Distribution of Respondents
The total amount of conducted interviews were 14, where participants from different backgrounds such as suppliers, governmental organizations and healthcare profes- sionals, all working with e-Health solutions, contributed with their views on the subject. The distribution of the interviewees work domain can be found in table 4.4. The interview questions can be found in Appendix A.
Supplier Governmental Organization Healthcare Professional
5 7 2
Table 4.2: Distribution of interview candidates.
The interviewee’s remain anonymous. However, below follows a short description of their background knowledge and work experience:
• Project leader with experience of eHealth and health informatics.
• Expert with experience of logistics within the healthcare organization.
• Expert with knowledge in health economics and drug administration.
• Representatives from two relevant medical technology suppliers.
• Expert with knowledge in medical technology systems and information manage- ment.
• Expert from SKL working on national level with structured patient-reported mea- surements.
• Project owner with experience of form service (Formul¨artj¨ansten in Swedish).
• Clinical modeller with experience of informatics.
• One representative from eHealth Authority working with digitalization on na- tional level.
• Two investigators from The Social Ministry.
• Two representatives from hospitals working with patients with chronic diseases.
• A representatives from Gothenburg University working with the project PROMIS.
Important to note is that most of the answers in the interviews were based on the participants personal views and thoughts, rather than their organization/employees approach towards the subject. This was confirmed by the majority of participants.
Text within quotation signs have been translated directly from Swedish to English, meaning that some expressions may be fairly distorted but without losing its mean- ing. Note that text within quotation signs are commentary, meaning, a reproduction of a content summation of the participants statements with own words.
4.1.3 Participatory Method
The results from the participatory method are presented below and divided into phase I and II. Each section presents the result generated under each phase. An example of the process of generating ideas on post-its during a interview session is presented in Figure 4.1.
Figure 4.1: Three notes created by a candidate during an interview.
Phase I
The participants generated new factors connected to PGHD during phase I. These are presented in table 4.2 and 4.3. The factors marked in bold were classified by the participants as the most important. The factors are connected to a keyword and domain based on suitable property explained by the participant. The keyword describes the factor while the domain places factor in the HTO perspective. The purpose of phase I was to detect important topics to be able to perform a thematic analysis.
Topics of Interest
Factor Domain
Clearer correlation Human-Organization Continuous measure-
ment
Human-Technology
Continuity Human
CROM Organization
Data management Technical Easier care flow -
Increased availabil- ity
Human
Follow health-chain Human
Infrastructure HTO
Information analy- sis
Organization
Large amount of data Technical Legal requirements Organization More quality at the
occasion
Human-Organization
Participation and Engagement
HTO
PREM - Patient Reported Experi- ence Measurement
Human-Organization
PROM - Patient Reported Outcome Management
Human-Organization
Purpose priority Human-Organization Quality of measure-
ments
Technical
Quality-assured equipment
HTO
Queuing times
(flow)
Organizational
Table 4.3: New factors created by interview participants.
Topics of Interest
Factor Domain
Responsibility for re- sults
Human-Organization
Queuing times
(equipment)
Organizational
Responsibility for re- sults
Human-Organization
Sign and verify Organization Simple equipment Technical
Security Technical
Semantics Technical
Statistics Human-Organization
”Too” much infor- mation
Technical- Organization
Transparency HTO
Trustworthy data HTO
Useful data Organizational
Value-stream- mapping
Organizational
Vision Human-Organization
Workload in the form of administra- tive work
Organizational
Table 4.4: Additional factors created by interview participants.
Phase II
Phase II is the validation part. All keywords except PROM, PREM and CROM (due to difficulty of understanding the words in general if not working with the subjects) were gathered on a paper and presented to the last three interview candidates after the interview questions were asked. The participants were asked to go through and analyze the keywords associated with PGHD created by previous candidates. Then they were asked to motivate which words they agreed with and if they had any objections. The possibility to add factors were also provided. All three participants agreed on that the generated factors were relevant to the subject. One candidate added an important factor, highlighted as darker grey, in table 4.3. The last two candidates focused more on explaining their experiences of the subject rather than
adding factors, as previous candidates had already covered many important ones.
4.1.4 Quantitative Questions
This section presents the outcome of the quantitative interview questions. The expected outcome from the questions was to explore knowledge, interest and current work of PGHD in different sectors. The result is presented in subsections below.
Knowledge and Interest
This section presents the answer to the interview question ”What is the general knowledge and interest in patient-generated health data in your work- place?”. The question is divided into two sub questions and the distribution of answers in the different work domains is presented in Figure 4.2 and 4.3. The Likert scale used to collect all the answers is presented in Figure 4.4.
Figure 4.2: Graphical distribution of general knowledge in PGHD.
Figure 4.3: Graphical distribution of general interest in PGHD.
Figure 4.4: Likert scale with four alternatives for each question.
Current Work
This section presents the answers to the questions ”Does your company work with patient-generated health data?” and ”How does your company or
organization work with patient-generated health data?”. The distribution of answers to the first question is presented in Figure 4.3.
Figure 4.5: Graphical distribution of general work with PGHD.
The three graphs in Figures 4.2, 4.3 and 4.5 present answers from 8 participants of a total 13. One participant chose not to respond. The remaining 4 participants answered the questions in the interviews but not according to the response options stated in the Likert scaled in later mailings. Thus, their answers are reflecting that their is both an interest and knowledge in PGHD and there is current work in pro- cess regarding PGHD at their workplace.
4.2 Data Analysis
This section presents the thematic analysis, the in-depth analysis of the factors and the answers to the qualitative interview questions in the data analysis. The thematic analysis consist of the factors presented in table 4.2, 4.3 and 4.4. The in-depth analysis consists of the factors occurrence in the various interview groups.
4.2.1 Thematic Analysis
The identified themes were created from the template analysis presented in phase I.
The 33 factors created during the interviews were divided into subcategories (light- blue rectangles) belonging to one of the four main themes (the grey circles) which are presented in figure 4.5.
Figure 4.6: A thematic analysis mapping of the factors from the template.
4.2.2 In-Depth Analysis of Factors
The thematic mapping in Figure 4.5 represent all factors created during the inter- views. Classification of the factors into the four themes was based on their charac- teristics. Participants discussed the effects on patient, healthcare and technology of PGHD when answering the interview questions, hence the above naming of themes
in Figure 4.5. The originated theme technology was split into two themes, digitized health data and equipment, because it provided the advantage of seeing a clear sep- aration of data processing in contradiction to measurement capacity and ease of use of the equipment. Factors created in the interviews differed depending on which interview group the participant belonged to. Therefore, an interesting aspect worth analyzing is the distribution of factors mentioned in the various interview groupings.
The main focus of the participants’ responses from the organization group was based on the factors in the ’patient behavior’ and ’healthcare organization’ themes. Their work and vision of promoting PGHD are highly motivated by government initiatives in line with the demand for access to care for citizens as well as facilitate medical work outcomes. Suppliers focused more on the tools och integration solutions mean- ing most factors produced by suppliers was gathered in the themes ’digitized health data’ and ’equipment’. The importance of creating the right tool for patients and healthcare to work with and meet the needs was at the center of the supplier’s responses. Healthcare professionals informed about tools used in healthcare today and effects that are predicted with the help of PGHD. The focus was mostly on the theme ’patient behaviour’ and ’healthcare organization’. However, all participants agreed that the reasons to why PGHD should be used can be attributed mainly to reduce medical burden in the long term and increase accessibility and quality of care for chronically ill patients. These four themes are interconnected and each part must fulfill its requirements to achieve a higher purpose.
4.2.3 Qualitative Questions
Below follows the answers to the qualitative interview questions asked during the interviews. They are presented under four subsections: work, value, healthcare flow and barriers.
Work
The following section presents the answers to the question ”How does your com- pany work with PGHD?” Some of the answers to how the participants workplace work with patient-generated health data is explained before the commentary marked with italics below.
One supplier described a feature in their product system which could identify and separate data created in healthcare with PGHD:
”There is a option to classify PGHD in our EMR-system. The system is able to distinguish between data generated in healthcare setting and the patients home en- vironment. The data is presented together, but it is possible to visualise where the data was generated and by whom.”
An expert answered that his department does not specifically work with projects on PGHD but stated that the organization finds it interesting and relevant:
”There are many research projects working on this topic. Initiatives are taken at Karolinska Institute where the patients are located at Karolinska Hospital.”
