Internet-Based Psychosocial Support

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1667

Internet-based Psychosocial Support

Design, Effects and User Experience in the Cancer Setting

ANNA HAUFFMAN

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6206 ISBN 978-91-513-0950-7 UPPSALA urn:nbn:se:uu:diva-408909 2020 Dissertation presented at Uppsala University to be publicly examined in H:son Holmdahlsalen, Uppsala Akademiska Sjukhus, ing 100, Uppsala, Friday, 12 June 2020 at 09:00 for the degree of Doctor of Philosophy (Faculty of Medicine). The examination will be conducted in Swedish. Faculty examiner: Professor Joakim Öhlén (The Sahlgrenska Academy, University of Gothenburg).

Abstract Hauffman, A. 2020. Internet-based Psychosocial Support. Design, Effects and User Experience in the Cancer Setting. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1667. 101 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0950-7.

Background and Aim Being diagnosed with cancer is often described as a major loss of control leading to severe psychological distress and symptoms of anxiety and depression can continue to affect the individual in the long term. The cancer and its treatment may influence all dimensions of health, thus the psychosocial support provided needs to be multifaceted and easy accessed. Internet-based interventions may be one way to provide such support, but evidence is limited. This thesis aimed to investigate the design, effects, and experiences of internet-based psychosocial support in cancer. Methods and Results Study 1 encompassed a co-creation development process resulting in the interactive support provided as the first step in an internet-based stepped care intervention (iCAN-DO). The effects of iCAN-DO were investigated in a randomised controlled trial, targeting individuals newly diagnosed with cancer and concurrent self-reported symptoms of anxiety and depression (according to the Hospital Anxiety and Depression Scale). Step 1 had a psycho-educative content involving self-care strategies and was available to the intervention group during the ten-month study period. Step 2 comprised a guided internet-based cognitive behavior therapy (iCBT) program and was offered those without improvement in anxiety and depression after using Step 1. The results showed that iCAN-DO improved symptoms of depression compared with standard care, while symptoms of anxiety were largely unaffected. Most participants used Step 1, while only a few used Step 2. In Study 2, aspects of usefulness, relevance, and usability in iCAN-DO were explored through qualitative interviews, analysed using content analysis. Results showed that standard healthcare did not meet the individuals' needs and iCAN-DO was used as complement, providing access to relevant, trustworthy information and support. Usability was affected by the perceived usefulness and ease of use of the intervention, as well as by the user´s circumstances in life and consequences of the cancer. The co-creation process in the development of Step 1 added relevance, but both steps 1 and 2 would have gained from being provided earlier, integrated into standard healthcare and more adaptable to the individual. Conclusion The thesis concluded that the internet-based intervention had positive effects on symptoms of depression in individuals newly diagnosed with cancer. Individuals with cancer experience several unmet needs in standard healthcare and since psycho-educative support including self-care advice seems feasible in this group, efforts are needed to incorporate internet- based support in regular oncology care. Since the intervention did not target all symptoms (i.e. anxiety) further research is needed on how to enhance efficacy and how to make iCBT more feasible for this group.

Keywords: Internet, eHealth, Nursing, Oncology, Cancer, Anxiety, Depression, Psychosocial Intervention, Technology-Based Interventions, Stepped Care

Anna Hauffman, Department of Immunology, Genetics and Pathology, Experimental and Clinical Oncology, Rudbecklaboratoriet, Uppsala University, SE-751 85 Uppsala, Sweden.

ISSN 1651-6206 ISBN 978-91-513-0950-7 urn:nbn:se:uu:diva-408909 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-408909)

To my son

As long as you live here You are the monarch of all you survey You own the whole earth

From Comet in Moominland by Tove Jansson

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Hauffman, A., Alfonsson, S., Mattsson, S., Forslund, M., Bill- Axelson, A., Nygren, P. Johansson, B. (2017). The development of a nurse-led Internet-based learning and self-care program for cancer patients with symptoms of anxiety and depression—A part of U-CARE. Cancer Nursing, 40(5), E9-E16.

II Hauffman A, Alfonsson S, Bill-Axelson A, Bergkvist L, For- slund M, Mattsson S, von Essen L, Nygren P, Igelström H, Jo- hansson B. Co-created Internet-based Stepped Care for Individ- uals with Cancer and Concurrent Symptoms of Anxiety and De- pression: Results from the U-CARE AdultCan Randomized Con- trolled Trial. Submitted.

III Hauffman, A., Alfonsson, S., Igelström, H., Johansson, B. Expe- riences of Internet-Based Stepped Care in Individuals With Can- cer and Concurrent Symptoms of Anxiety and Depression: Qual- itative Exploration Conducted Alongside the U-CARE AdultCan Randomized Controlled Trial. Journal of Medical Internet Re- search, 22 (3):e16547.

IV Igelström, H. Hauffman, A. Alfonsson, S. Sjöström, J. Cajander, Å. Johansson, B. Delivery, design, and structure of an internet- based stepped care intervention targeting individuals with cancer and concurrent symptoms of anxiety or depression: a qualitative study of user experiences from U-CARE AdultCan. Journal of Medical Internet Research; ahead of print.

Reprints were made with permission from the respective publishers.

Contents

Introduction ...... 11 The impact of cancer ...... 12 Receiving a cancer diagnosis ...... 12 Receiving cancer treatment ...... 13 Common impairments following cancer and cancer treatment ...... 13 Development of anxiety and depression in cancer ...... 15 Identification and assessment of symptoms ...... 17 Patient-reported outcomes ...... 18 Patient-reported outcome measures in the cancer context ...... 19 Psychosocial and psychological interventions promoting health in individuals with cancer ...... 21 Psychosocial support in regular healthcare during cancer treatment in Sweden ...... 22 Psychoeducation and self-efficacy ...... 23 Self-management and self-care ...... 24 Cognitive behavioural therapy ...... 25 Stepped care ...... 26 The internet as a tool to improve health ...... 26 Rationale for the thesis ...... 30 Overall and specific aims ...... 31 Methods ...... 32 Design ...... 32 The U-CARE Portal ...... 32 Participants and settings ...... 33 Study 1 (Papers I–II) ...... 33 The intervention ...... 36 Study 2 (Papers III–IV) ...... 43 Ethical considerations (Studies 1 and 2) ...... 45 Results ...... 48 The development (Paper I) ...... 48 Initial test of acceptance ...... 50 The AdultCan randomised controlled trial (Paper II) ...... 50 Utilisation ...... 52 Perceived benefits ...... 53

Anxiety and depression ...... 53 Health-related quality of life and post-traumatic stress ...... 56 User experiences of relevance and benefit (Paper III) ...... 58 Gaining knowledge and support but wanting more personalisation .... 59 A feeling of safety that was needed earlier, and the use of peer- support ...... 60 A complement to standard care ...... 61 User experience of delivery, design, and structure (Paper IV) ...... 63 User experience in the context of cancer ...... 63 Technical struggles require patience and troubleshooting ...... 64 Appealing and usable, but rather simple ...... 65 Discussion ...... 67 Main findings ...... 67 The development of interactive support (Paper I) ...... 67 The effects of iCAN-DO (Paper II) ...... 69 Effects on anxiety and depression ...... 69 Effects on HRQoL and post-traumatic stress ...... 71 Did the intervention happen? ...... 71 Perceived relevance, benefits, and aspects of usability (Papers III–IV) .. 72 User experience of relevance and benefit ...... 72 User experiences concerning usability ...... 75 Methodological discussion ...... 76 Study 1 ...... 76 Study 2 ...... 79 Conclusions ...... 80 Clinical implications and further research ...... 81 Svensk sammanfattning (Swedish summary) ...... 82 Studie 1 ...... 82 Delarbete I ...... 82 Delarbete II ...... 83 Studie 2 ...... 84 Delarbete III-IV ...... 84 Slutsatser ...... 85 Acknowledgements ...... 86 References ...... 88

Abbreviations

BPS Biopsychosocial model CBT Cognitive behavioural therapy EORTC QLQ-C30 European Organisation for Research and Treat- ment of Cancer Quality of Life Questionnaire- Core 30 EROS The Environmental Reward Observation Scale FACIT-F The Functional Assessment of Chronic Illness Therapy-Fatigue FAQ Frequently asked questions FCR Fear of cancer recurrence HADS Hospital Anxiety and Depression Scale HRQoL Health-related quality of life IAS Interactive support iCBT Internet-delivered cognitive behaviour therapy IHCA Interactive health communication applications ISI The Insomnia Severity Index MADRS-S Self-reported Montgomery Åsberg Depression Rating Scale MCID Minimal clinically important difference PCL-C The Post-Traumatic Stress Disorder Checklist- Civilian version PROM Patient-reported outcome measure PTG Post-traumatic growth RCT Randomised controlled trial S-CDNT Self-care deficit theory of nursing SCG Standard care group SCT Social cognitive theory STAI-S The State-Trait Anxiety Inventory (State) TAM The technology acceptance model U-CARE Uppsala University Psychosocial Care programme WHO World Health Organization

Introduction

The terms disease, illness, and health are interconnected. “Disease” is used to describe the condition that usually becomes apparent in connection with a diagnosis and is the term that often represents the suffering in a biomedical model. This has long been the dominating model on which healthcare systems have been built. However, it is possible to be diagnosed with a disease without the experience of suffering from illness (1); likewise, it is possible to be declared free from disease but still suffer consequences affecting health (2). Thus, the experience of health is individual, starting from the subjective experience of the meaning of quality of life. Disease, illness, and health are also subject to constant change, and can be looked upon as a continuum in which all individuals are continually moving. The changes can be looked upon across different timespans, e.g., throughout life or by the hour (3). A model with great impact on modern nursing is the biopsychosocial model (BPS) (4). In the BPS, the medical or biological dimension must be accompanied by social, psychological, and behavioural dimensions to explain the state of health. The BPS is the model most in line with the World Health Organization’s (5) definition of health as a state affected by mental, physical, and social factors, not merely the absence of disease. Within nursing, the components “person” and “environment” are also conceptualised (6); the person is seen with needs springing from the body, the mind, and the soul and– in one way or another – engages in human relationships to find meaning. The environment includes the “real” structures surrounding a person, but also what goes on in their consciousness. Thus, to approach an understanding of the impact of a disease, numerous aspects must be taken into account. This is implemented in the nursing competence (7), e.g., by always considering the individuals life situation and own values, including the individual as a full member of the team and by working together with other professions to ensure a comprehensive approach. In short, the individual cannot be reduced to a disease and even if improved health is the desired outcome, healthy dimensions will already exist, regardless of any disease.

As a result, a disease will not entail the same experience for any two individuals. Aside from our different biological conditions, we have all had different lives, shaping us into who we are. A person diagnosed with cancer is a single traveller through that experience and the journey will not be experienced in the same way by each person, when the physical, psychological, social, and

11 existential impact of the disease merges into that person’s life (8). Neverthe- less, receiving a cancer diagnosis is commonly described as a major loss of control leading to severe psychological distress (9). For most people, the experience of being diagnosed and treated for cancer will be completely new, and therefore knowledge on how to handle the situation will be lacking. There- fore, learning is often described as a key factor in order for individuals to gain control, and, consequently, nurses within oncology describe teaching as a crucial element in their professional practice (10). The main purpose of such teaching is to assist the individual and their next-of-kin in being able to exist in the new situation. This can involve helping them understand the illness, the symptoms, the healthcare system, their own reactions or something completely different, depending on their needs. Most individuals with cancer will be cared for in an outpatient setting, which of course may be perceived as positive, as the routines of everyday life may be less affected. However, the cost is that individuals are often left to manage their situation relatively alone, with the healthcare system at a distance, facing major challenges of its own. If the individual is to get relevant support, this may require a strong power of initiative, which for natural reasons is not always present. Further, the healthcare system is fractured into specialties, where individuals with cancer and chronic illnesses have trouble fitting in. What the individual may experience as a single problem may require multiple visits to different specialties and disciplines, and it is not uncommon that people with complex diseases such as cancer describe several unmet needs (11, 12). In order to bridge this gap, it is vital that we investigate new comprehensive methods to provide relevant support and knowledge to help individuals manage their entire situation before, during, and after the cancer and its treatment. As many individuals with cancer are already using the internet when looking for cancer information (13, 14), this serves as a natural arena to study within this field. This thesis will explore support and knowledge delivered via the internet; how it could be designed, what effects it may have, and how it may be perceived.

The impact of cancer Receiving a cancer diagnosis More than one in three people living in Sweden will receive a cancer diagnosis during their lifetime. Breast cancer is the most common diagnosis among women and prostate cancer is most common among men, followed by skin and colorectal cancer for both men and women. Many of those affected have suspicions, concerns, or realisations about their disease already in the pre-diagnostic phase. Others may be more unprepared for the news of their cancer. Thus, the starting point can vary dramatically. Still, most of us know someone affected by cancer and are aware that cancer may lead to death. Even if the risk of death due to cancer has decreased, cancer is still the most common

12 cause of death for people under the age of 80 years in Sweden; globally one in six deaths is due to cancer (15, 16). Thus, the diagnosis may be perceived as forcing a confrontation with perceptions of our own mortality and is often described as a highly stressful event, involving feelings of fear and helplessness. The time close to the diagnosis is often filled with uncertainty, not knowing what to expect, neither for the next few weeks nor for the rest of one’s life. In addition to shock, numbness, and sadness, symptoms of anxiety and depression may be part of the initial “natural” reaction. These may subside but can also be consistent (17-19). While the meaning and impact of a cancer diagnosis is subjective, it could be looked upon as a crisis where the ensuing reactions depend on the individual’s earlier experiences, vulnerability, relations, and the perceived level of threat (20). The occasion when the diagnosis is communicated has been shown to be highly important and if possible, the person should be accompanied by someone close to them. The perceived emotional support at, and directly following, the occasion when the diagnosis is communicated may predict the affected person’s sense of hope (21). Feelings of hopelessness are, in turn, closely associated with the vicious cyclic development of depressive symptoms (22).

Receiving cancer treatment The treatment phase can impose a clear structure on the individual’s existence, in which they may find some peace. The time of waiting is over and something that may be perceived as positive is happening. Moreover, during the treatment phase, there are regular possibilities to get support from healthcare staff. The most common medical treatments are surgery, chemotherapy, radiation, and hormone treatment, depending on the type of cancer. The medical treatments are often combined (e.g., surgery followed by chemotherapy and radiation). Additional targeted treatments, e.g., with antibodies, may also be administered. These treatments are fundamental to survival, but often have side effects, as they can be toxic, damaging healthy tissue or altering the body’s original function. Most side effects are transient, while some may be long- lasting or chronic (23, 24). Thus, receiving treatment is not always purely positive, as the body will most likely react or change. It has therefore been suggested that an individual with cancer may not pass through only one crisis, but several, as each treatment and its effects can be perceived as a new one (25). In addition to treatment of the tumour, drugs targeting symptoms, e.g., nausea, anxiety, depression, and sleep disorders, may be prescribed by the physician.

Common impairments following cancer and cancer treatment When evaluating the impact of disease and treatment, health-related quality of life (HRQoL) is often used as a measure. HRQoL may be seen as the subjec-

13 tive perception of quality of life in relation to the impact of a disease, encompassing physical, psychological, social, and functional dimensions (5, 26). A Swedish population-based study (27) has found that disease and physical in- activity are the most important covariates of poor HRQoL. Cancer and its treatment may cause a wide variety of symptoms and influence all dimensions of HRQoL (27). In a review of symptom prevalence among people receiving cancer treatment (28), a total of 47 different symptoms was reported, fatigue, insomnia, and pain being the most common. Some of these symptoms may be caused by the disease itself, while some may be consequences of treatment, or simply appear as a result of another symptom. It is assumed that symptoms have a catalytic effect, e.g., the symptom of pain has been shown to directly influence fatigue, but pain may also lead to sleep disturbances, which in turn affects fatigue. Pain may also be perceived as considerably worse in the presence of fatigue (29, 30). These associations between symptoms have given rise to the concept of symptom clusters, i.e., symptoms that often occur together. Further, symptoms that occur as a result of cancer treatment are divided into early and late symptoms. The early symptoms are often somewhat expected, e.g., hair loss, nausea, fatigue, and irritation of the skin and mucous membranes. Most of the early symptoms are transient and subside upon completion of treatment. The late symptoms may not be as expected or distinct, often involving loss of sensory perceptions and memory. The late symptoms have been shown to have a large impact on everyday life, with negative effects on HRQoL (31).

In addition to symptoms, cancer may lead to a lot of changes, including to relationships, personal economy, and functional abilities. Limited physical functioning is common; in a population-based study, 53% of those who had cancer reported physical functional limitations compared with 21% of those with no history of cancer (32). These limitations may appear before, during, or after cancer treatment and have been suggested to have the highest impact on those who are below 65 years of age (33). The impact on family members is well known and the prevalence of psychological problems and the use of healthcare are higher among partners of those with cancer than among healthy controls (19, 34). The outpatient treatment settings have increased the responsibilities of next-of-kin, who often find themselves in a caregiving role (35). Still, in the latest national patient inquiry, only 50% of individuals with cancer stated that their next-of-kin had been offered any emotional support from regular healthcare (36).

No matter what is affecting the individual, it is important to grasp the entirety of the situation. All symptoms, impairments, and changes may be reported as mild to moderate, but it must be acknowledged that it is their total impact that ultimately affects the individual’s health (37). As a man with cancer once simply expressed it: “It is important that you understand that this will all land

14 in the same person.” Also, when meeting a person with several physical symptoms, it is more likely that there are other problems present, such as symptoms of anxiety and depression (38). Thus, the support provided should have a comprehensive approach, targeting the situation following the cancer as a whole rather than focusing on isolated symptoms.

Development of anxiety and depression in cancer According to what we know, most individuals diagnosed with cancer will not experience long-term psychological problems, such as anxiety and depression, but some do experience persisting symptoms. The constructs of anxiety and depression encompass cognitions (I won’t manage it/I’m worthless), feelings (fear/guilt), physical reactions (restlessness/breathing difficulties), and behaviours (irritability/withdrawal) (39). Both anxiety and depression may be seen as normal reactions to an overwhelming life situation where the individual experiences severe stress, helplessness, and loss of control. Biological factors may partly explain the development of symptoms, but psychological and social factors must be recognised as highly important. For example, the fear and vulnerability associated with the real danger inherent to cancer is most likely a trigger for anxiety in the early phases of the disease. In a longer perspective, the fear of recurrence (FOR) (in the cancer context also referred to as FCR; fear of cancer recurrence) is commonly described as a trigger (40). Depression symptoms may be fuelled by feelings of helplessness and loss of control and meaning. Symptoms can also grow stronger due to the changed behaviours resulting from the new feelings. Further, the individual’s own perception of their abilities to cope with the situation (self-efficacy) has been suggested as an important factor (41).

A partial explanation may be offered by the stress-vulnerability-model (42), which encompasses both biological and acquired factors that play an important role in the development of anxiety and depression. Thus, the individual’s vulnerability will be partly congenital, but past life event stressors, such as grief, fear, isolation, and stress, are at least as important. This can be compared to a weight on a string; it initially holds the weight, but when/if additional weight (such as cancer) is placed on the string, it will eventually break. The stress-vulnerability model may also be illustrated as a bucket (Figure 1), where the individual coping mechanisms and positive environmental exchange serve as outlets, helping to ensure the bucket does not overflow (43).

Figure 1. Stress-vulnerability model. Developed from Brabban and Turkington (2002).

Anxiety and depression during the cancer trajectory The prevalence of anxiety and depression in cancer is studied with a variety of methods in many different settings. Previous research indicates that symptoms of anxiety, depression, or distress in cancer may be present before, during, and after treatment in 10–60% of patients (18, 44-46). The development of anxiety and depression symptoms in cancer has previously been shown to depend on, e.g., diagnosis, treatment (specific neuropsychiatric effects), gender, age, prognosis, comorbidity, existential distress, socioeconomic position, and previous life experiences (47-50). As stated, some findings are reported as symptoms of anxiety and depression, while some researchers use the term ‘psychological distress’ (51). Distress may be used differently, describing un- pleasant emotional experiences, such as strain, stress, fear, or sadness. It is also used to describe a combination of anxiety and depression, e.g., when reporting the total score on the Hospital Anxiety and Depression Scale (HADS) (52), where the categories non-distress and distress are recommended (53). The occurrence of both anxiety and depression symptoms is very common in cancer, with two thirds of individuals with depression also reporting symptoms of anxiety (54). A recent study (55) comparing levels of anxiety and depression in cancer with those in a general population found that individuals with cancer suffered from more anxiety than the general population, regardless of age. Women and younger individuals with cancer showed the highest levels of anxiety. As for depression, men with cancer did not differ much from

16 the general population, while especially younger women with cancer reported more depressive symptoms than the general population. Other studies have also examined which people are most likely to be affected, and when. In a study of more than 10,000 individuals with different cancer diagnosis, 19% showed clinical levels and 22% subclinical levels of anxiety and depression, across cancer types. Women and younger individuals showed higher rates of both anxiety and depression following diagnosis than men and individuals above the age of 50 years (56). However, the relevant measures when studying depression in men with cancer may be different (57); it has previously been suggested that men with symptoms of depression can be easier to detect through structured clinical interviews than via questionnaires (58). A systematic review of symptoms of anxiety and depression in prostate cancer patients identified prevalence rates pre-treatment, on-treatment, and post-treatment, suggesting that both anxiety and depression symptoms decreased during the treatment period and then increased after completion of treatment (59). Con- versely, a meta-analysis containing mixed diagnoses found the highest prevalence of depression during active treatment (60).

Nevertheless, symptoms have been shown to be long-lasting. A Danish nationwide study found an increased risk of being admitted to hospital due to depression among those diagnosed with cancer. This risk was at the highest one year after diagnosis, but persisted for ten years (61). Further, depressive symptoms have previously been described as a predictor for elevated mortality in cancer (62), increasing the risk of suicide (63, 64), generating higher costs (65), and making the individual less adherent to cancer treatment (66, 67). Still, most individuals with these problems remain undetected and untreated (68).

Identification and assessment of symptoms When identifying symptoms in clinical cancer care, the meeting and conversation with the individual is usually the core, sometimes in combination with measurement instruments, depending on the preferences of the nurse or clinician. However, problems within psychological dimensions following cancer are not always as obvious as physical symptoms, and standard healthcare is having problems with their identification (69, 70). In a study of the use of validated questionnaires to detect mood disorders in cancer care, 10% or less of the responding clinicians used questionnaires alongside conversation (71). On the other hand, within psychology and psychiatric care, structured clinical interviews are often used in combination with validated questionnaires. The questionnaires may be patient-reported, i.e., administered by the individual, or clinician-reported, i.e., administered by the healthcare professional.

17 Patient-reported outcomes Patient-reported outcomes (PROs) can be defined as the outcomes that are complementary to disease-focused outcomes (e.g., survival), covering the individual’s own perceptions of the impact of disease and treatment, without interpretation by someone else. The growing recognition of the BPS and real- isation of the value of overall wellbeing in individuals with diseases have increased the use of such outcomes (72). PROs are often assessed using validated questionnaires, resulting in patient-reported outcome measures (PROMs) (73), classified by length and symptoms being assessed. Tradition- ally, PROMs are used as paper forms, but many questionnaires are currently being validated and developed to be internet-administered: ePROMs. Several barriers to using PROMs in clinical care have been reported (e.g., logistical and social) (74), and ePROMs may avoid these barriers. When using a PROM as a screening tool, the purpose is to identify the individuals who are experi- encing symptoms or are at risk of developing symptoms. This enables offering those identified, e.g., information, further investigations, and/or treatment, (such identification falls within the contact nurses’ responsibilities). The PROM used should therefore be as accurate as possible to be useful. The PROM needs validity, i.e., needs to measure what it is intended to measure, and reliability, i.e., show consistency in repeated measures (75). It has been suggested that at least one full report on the development of a test and one on the use of the test are needed for evaluation of its psychometric properties (76). The accuracy of PROMs as screening tools is described through its sensitivity and specificity. If a PROM were to have 100% sensitivity, it would correctly identify all individuals with the relevant symptoms. With a 100% specificity, the results would be negative for all individuals without the relevant symptoms (77). However, sensitivity is often prioritised in screening procedures.

Regarding interpretations, guideline values are often suggested for each PROM. In addition, minimal clinically important differences (MCIDs) (78) within and between groups may be available to enable interpretation of results. An MCID can be defined as the smallest change in outcome that is valuable to the individual and that would lead to considering a change in management of the problem/symptom investigated (79). To determine an MCID, a combination of distribution- and anchor-based methods are often used, together with the Delphi method (80).

Interviewing In addition to letting individuals self-report symptoms via questionnaires, clinical interviews may be used. These interviews are structured with open- ended questions and one of the advantages is that the individual’s life context may be captured alongside functional abilities and symptoms. The research

18 context may also include interviews to explore individuals’ subjective views. Lewin et al. (81) specifically described the use of qualitative methods such as interviews, alongside randomised controlled trials (RCTs), to gain better understanding of aspects surrounding an intervention. Thus, qualitative inquiries are recommended before a trial to enhance development or generate hypothe- ses, during a trial to explore the delivery and participants’ responses, and after a trial to explain the findings.

Patient-reported outcome measures in the cancer context As described, cancer affects the individual multidimensionally; thus, there are many PROMs that could be used in the cancer context. To avoid measurement inconsistency, relevant domains and measurements for the cancer context have been suggested following a Delphi consensus process (PROMS-Cancer Core) (73). In the first scoping review, more than 1,500 citations were included. The agreed domains were HRQoL, Physical health, Emotional health, and Social health, which were further narrowed down into subdomains, dimensions, and, lastly, measures. The measures used in this thesis fall within these recommendations in several, but not all, cases. The consensus process concluded that further research regarding best measures in some domains were still needed. While the PROMS-Cancer Core recommendations are not complete, they provide the best standardisation found within the literature.

Quality of life HRQoL was identified as a highly meaningful outcome by PROMS-Cancer Core and the main measure recommended for assessing the impact of disease and treatment on quality of life was the European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire Core 30 (QLQ-C30) (82). It aims to capture the subjective whole of the individual’s situation in relation to the cancer and its treatment. The QLQ-C30 includes functional (e.g., role functioning), physical (e.g., nausea), psychological (e.g., worrying), and social (e.g., relationships) dimensions (82). In addition, there are diagnosis-specific modules for several diagnoses, e.g., breast cancer (QLQ-BR29), prostate cancer (QLQ-PR25), and colorectal cancer (QLQ- C29), focusing on specific symptoms. An alternative measure of HRQoL recommended is the Functional Assessment of Cancer Therapy-General (FACT- G) (83). There are several differences between the QLQ-C30 and the FACT- G. For example, the FACT-G contains subscales of “wellbeing” in different dimensions instead of “functioning” scales and uses statements instead of questions. A previous systematic review (84) comparing both instruments could not recommend the QLQ-C30 over the FACT-G, but instead suggested close review of both measures in relation to the research objectives. Moreover, while both the QLQ-C30 and the FACT-G are multidimensional tools, additional instruments may be needed for assessment of specific symptoms (85).

19 Physical health Recommendations within the physical health domain include specific assessments of symptoms, physical function, sexual function, and cognitive function. Since the symptomatology may vary between diagnosis, this is the largest domain, and includes multiple instruments that may differ in relevance to different individuals. However, two common, multi-factorial and interacting symptoms across cancer types are insomnia and cancer-related fatigue (CRF). They are considered to form a symptom cluster with depression (86), making them especially important to monitor. In this domain, two outcome measures for insomnia are recommended, one of them being the Insomnia Severity In- dex (ISI) (87). This is a multidimensional, short, self-reported measure of insomnia. A recent study of 100 individuals, 1–2 years after cancer treatment, found the ISI (English version) to be a reliable measure in that population (88). Five measures are recommended to assess CRF, e.g., the Functional Assess- ment of Chronic Illness Therapy-Fatigue (FACIT-F) (89). This is a short tool, originally developed for individuals with cancer. The FACIT-F may be used as an extension to the FACT-F or as a stand-alone instrument.

Emotional health The emotional health domain includes the dimensions anxiety, depression, and psychological adjustment. The main measure recommended for assessment of both anxiety and depression is the short instrument HADS (52). It has previously been shown to be valid (38, 90) and is the most commonly used scale when screening for symptoms of anxiety and depression among individuals with cancer (91). One of the advantages of this measure is the account taken of the underlying disease, as disease symptoms may otherwise confound the symptoms of anxiety and depression. The questionnaire may be divided into two subscales: HADS-A (for anxiety) and HADS-D (for depression). They may also be interpreted together (HADS-T), with distress used as the outcome. Recommendations within this domain did not include distress as a separate outcome but focused on the specific symptoms of anxiety and depression. Another specific measure recommended is the Spielberger State-Trait Anxiety Inventory (STAI) (92). This is a short measure containing two modules, one assessing anxiety at the moment (s = state), and another the more permanent situation (t = trait). Each module consists of 20 statements, e.g., “I feel upset” or “I worry about conceivable accidents”.

Some interesting considerations were noted within this domain, e.g., self-efficacy was highlighted as an important outcome, but no outcome measure specific to cancer could be identified. Therefore, it was excluded. Further, anger and substance abuse were also considered as important measures, but the panel could not reach consensus and such measures were excluded. Still, anger is a

20 known specific symptom of depression in men with cancer (57) and assessment could therefore be motivated in such populations. Another reflection is that no measure for cancer-specific stress is recommended. Being diagnosed with a life-threatening disease has formerly been put in the same class as being raped or surviving war or a tsunami. However, a change in the classification means that the DSM-V does not include cancer as a traumatic event causing post-traumatic stress disorder (PTSD) (93). Formerly, the post-traumatic stress disorder checklist-civilian version (PCL-C) (94) was commonly used to assess PTSD symptoms in cancer. An alternative, commonly used measure of cancer-specific stress or trauma is the Impact of Events Scale (95), measuring the subjective impact of a specific event. It is a short measure consisting of 15 items on a four-point Likert scale, with a higher score indicating higher levels of cancer-specific stress. Meanwhile, post-traumatic growth (PTG) has been observed among some individuals treated for cancer. This is commonly described as the positive change that may occur following a life crisis. In individuals with cancer, results of PTG have been reported, regarding for instance increased self-perceived inner strength, greater appreciation of life, improved relationships, and more positive health behaviours (93). Although cancer-specific stress is not handled within the recommendations, PTG is mentioned as a relevant outcome.

Social health The social health domain, including functioning and relationships, is the smallest of the domains, encompassing just two measures: the Medical Out- comes Study Social Support Survey (96) and the Psychosocial Adjustment to Illness Scale (97). The recommendations suggest further investigation of such measures. Further, social costs were highlighted as important, but were not seen as falling within the scope of the recommendations.

Psychosocial and psychological interventions promoting health in individuals with cancer The term “psychosocial” has a close connection to the BPS, and encompasses the physical, psychological, and social factors affecting an individual’s health (98). The psychosocial aspects of cancer include interventions that besides psychosocial may be described as psychoeducational, psychological, psycho- therapeutic, or psycho-oncological. A Cochrane review (99) have assessed psychosocial interventions regarding the effects on anxiety, depression, and distress in newly diagnosed individuals. Results suggested that such interventions added value to cancer care, especially those concerning nurse-delivered psychoeducational efforts. Several problems have been identified in the field of psychosocial interventions in cancer. An early systematic review (100)

21 found moderate effects in symptoms of anxiety, but no effects in symptoms of depression. However, a later review (101) concluded that studies in the field were of such low quality that the positive results on emotional outcomes had to be interpreted with caution. The Cochrane review previously mentioned (99) concluded that most interventions targeted participants with a cancer diagnosis, rather than considering their perceived need for support. A more recent systematic review and meta-analysis (102) found flaws in the descriptions of interventions, non-evidence based interventions, and a lack of inclusion criteria for participants. The conclusions drawn were that testing of interventions, screening, and stepped-care models should be included in future studies.

Psychosocial support in regular healthcare during cancer treatment in Sweden In addition to the basic support from regular healthcare staff, psychosocial support may be provided by self-referral to a counsellor or a hospital church. In addition, there may be psychosocial initiatives or projects that vary between regions and hospitals. Moreover, an increasing number of diagnostic groups include a contact nurse who shall provide basic psychosocial support, in accordance with the national cancer strategy (103). However, it remains unclear to what extent psychosocial support has been implemented in the role of the contact nurse, and there seem to be regional differences regarding the interpretation of the contact nurse’s role. In November 2019, the Swedish cancer centres published an elucidation regarding the assignment of the contact nurse, highlighting the importance of providing psychosocial support at diagnosis as well as during and after treatment for cancer (104). Lastly, an individual’s own perceptions of the care and support that they receive during their illness is of utmost importance. As stated earlier, aside from surgery and sudden acute care needs, cancer treatment takes place in an outpatient setting, with short visits to the hospital. The contents of these visits are usually determined in advance by the healthcare system as diagnostic procedures or treatment to be performed; this may be seen as a result of the ruling biomedical model. Some may find it easy to adapt to this situation, while others may find it a major loss of integrity (25). The experience of being in a caring process has previously been described as humanising or de-humanising for the individual (105). Cru- cial for humanising care is that the individual perceives that caregivers communicate and work with them to retain their dignity and overcome their suffering. The opposite may occur when the individual feels that they have to fight the healthcare system to get help and support.

22 Cancer rehabilitation The term cancer rehabilitation is used increasingly in the Swedish cancer care context and encompasses all the efforts to reduce the long-term impact of cancer, including physical, social, existential, and mental rehabilitation efforts. Under the national cancer strategy (106), cancer rehabilitation should be offered to all patients throughout the entire cancer trajectory. The starting point of the rehabilitation process is the time of diagnosis. As a result, cancer rehabilitation is a team effort, involving several professions. The individual rehabilitation needs are subjective, but also dependent on the outcome of treatment. In the best scenarios, treatment will lead to remission, from a biomedical perspective, 65% will be cured of their cancer (107). Despite being cured, individuals may continue to experience consequences of the cancer and its treatment for a long time, sometimes the rest of their lives (108). As previously stated, a person may undergo medical treatment, be declared fully treated and cancer-free, but still suffer consequences affecting health (2). In addition, next-of-kin should always be included in the rehabilitation plans. If remission does not materialise, or a relapse occurs, the needs will naturally look different and involve acceptance of cancer as a part of the remaining life. However, it remains unclear to what extent the national recommendations on cancer rehabilitation are or will be implemented. In addition, there are several known barriers to successful cancer rehabilitation (109), some within the structure of healthcare, some within the individual, and some resulting from the stress caused by the disease.

A large Swedish study has recently investigated the interest among individuals with cancer to participate in rehabilitation (110). The rehabilitation activities most preferred by the respondents included psychoeducation, open lectures on cancer, support provided by a social worker, physical activity assisted by a physiotherapist, and internet-based support. Results also suggested that the interest to participate in different rehabilitation activities varied by level of education, gender, age, and time since diagnosis. This suggests that the rehabilitation activities offered need to be multifaceted and offered at different times during the cancer trajectory.

Psychoeducation and self-efficacy As stated, psychoeducational interventions has been suggested to be effective (99) in the cancer context. Psychoeducation is one of the tools within cognitive behaviour therapy (CBT) (111, 112). One of its basic assumptions is that a situation, symptom, or problem will be more bearable if normalised, explained, and understood. The main components within psychoeducation involve providing general information in combination with an explanatory model. This makes it highly suitable to empower newly diagnosed individuals.

23 The increased knowledge may in turn affect self-efficacy, i.e., the individual’s own perception of their ability to perform a task. Self-efficacy is one of the original constructs of the social learning theory described by Bandura (41), which has now evolved into social cognitive theory (SCT) (6). Therein, self- efficacy is described as having the potential to affect behaviours. These behaviours will in turn affect the surrounding environment, and the environment will respond. The response may then affect the individual’s thoughts and behaviours. Self-efficacy expectations may determine how long an individual will “persist” when facing difficulties in life. The experiences of an individual who persists even when feeling threatened will further reinforce the own experience of efficacy and vice versa. In addition to psychoeducation, interventions may encompass several other pathways to increase an individual’s practical, emotional, and cognitive capacities.

Self-management and self-care There is a lack of consensus regarding the definition of self-management, but existing definitions involve management of own health with regard to medical aspects, psychological consequences, and new life roles over time (113). This again implies the need for a comprehensive approach in interventions. A recent systematic review (114) on self-management targeting individuals with cancer found a great heterogenicity among interventions regarding length, delivery, and underlying theories and components. The same conclusions were drawn in an earlier review (115). Thus, it was not possible to recommend which type of self-management should be integrated as a part of standard care, or even if it should be integrated at all. Linden et al. (116) concluded that the lacking evidence was a product of not including individuals with defined problems in psychosocial trials in cancer, i.e., not using screening procedures. Get- ting rid of the floor effect would result in higher treatment effects. Another problem in deciding what interventions are the most meaningful may be poor descriptions of intervention contents and how they are believed to exert an effect.

A concept very close to self-management is self-care (an exact delineation between the two could not be established in the literature). Definitions of self- care also vary, but an overall definition would concern the individual’s own resources and abilities to perform actions to maintain and improve health. The self-care concept may be expanded to include actions to promote the wellbeing of next-of-kin, e.g., children and family members (117). The nursing the- orist most strongly associated with the concept of self-care is Dorothea Orem. Her Self-care Deficit Nursing Theory (S-CDNT) consists of three major parts (6, 118). One is the theory of self-care, describing the constructs of self-care, including the concepts of self-care demand and self-care agency. The second is the theory of self-care deficit, describing the imbalance that occurs when

24 the self-care demand exceeds the individual’s capacity (the self-care agency), and the third is the theory of nursing systems, describing the nurse’s role as educative and supportive, with a conscious agenda to enable self-care. The S- CDNT has previously been found useful in interventions within burn care (119), pelvic floor rehabilitation (120), cystic fibrosis (121), and advanced heart failure (122), as well as in the cancer context (123). The last of these studies used Orem’s model to determine predictors of self-care behaviour. Re- sults suggested that individuals with higher levels of anxiety, less social support, and higher education were most likely to perform self-care.

The constructs within SCT have been used together with Orem’s theories in several interventions. Desbiens et al. (124) investigated the linkage between the two and found 14 articles using the components in various ways, e.g., with the SCT as a basic conditioning, a mediator, or a component of the self-care agency. Richard et al. (125) explored the concepts surrounding self-care and found two constructs of importance to nursing, beyond self-management, self- care, and self-efficacy: self-monitoring and symptom management. Self-monitoring is described as the individual’s ability to be aware of (e.g., through measurement) when to either perform an action themselves or to contact healthcare providers. Symptom management is described as a combination of the individual’s awareness of and response to changes in own health. In this review, self-efficacy was described as having the potential to weaken or strengthen the other factors mentioned. Thus, it can be presumed that self- efficacy can influence an individual’s self-care behaviour.

Cognitive behavioural therapy As stated, anxiety and depression are related to the interaction of emotions, cognitions, and behaviours. CBT is a prominent therapy focusing on these constructs with roots in learning theory (111, 112). Learning and behaviour are central aspects of CBT and its core principles include the assumption that psychological problems are partly based on unhelpful ways of thinking and unhelpful behaviours. Since our feelings and cognitions may be hard – or even impossible – to change consciously, the method focuses on “going in through the back door” and altering behaviours that in turn will influence feelings and cognitions. It is a therapy focused on the present and future, including a thorough analysis of situations involving “ineffective” thoughts and behaviours, to support the activation of new ones. Thus, the main focus is not the problematic behaviours, but rather the creation of new ones, and consequently self- efficacy is an important factor within CBT. According to the Swedish Agency for Health Technology Assessment and Assessment of Social Services, CBT has the highest strength of evidence for treatment of depressive symptoms (126). While the scientific support in cancer populations is not as extensive, CBT has been shown strong positive effects regarding specific symptoms of

25 cancer, such as insomnia and fatigue (127, 128). In Sweden, CBT is also recommended as part of the rehabilitation in breast cancer for example (129), but may not be as effective in individuals with advanced cancer (130). The design of studies involving CBT must be considered carefully, like in the case of studies of self-management, as the treatment effects will most likely be absent in individuals without self-reported symptoms. Further, CBT alongside cancer treatment is recommended to incorporate the cancer care i.e. “adjuvant psychological treatment” (131).

Stepped care The current challenges in the public economy must be considered when planning new interventions. One model that serves to increase cost-effectiveness in psychological therapies is the stepped care model (132). Stepped care involves treatment given at differing intensities and is based on the assumption that low-intensity interventions directed at the right population will provide health gains similar to those of more resource-intensive interventions. If the low-intensity intervention does not lead to improvement, the next step of intensity is applied. In the example below, self-care and self-management con- stitute the first step (see Figure 2). The stepped care model may be appropriate in the cancer context and is currently being evaluated, e.g., in a nurse-led internet-based intervention, to enable a return to work after cancer (133).

Figure 2. A simple two-stepped model developed from Bower and Gilbody (2005).

The internet as a tool to improve health It is estimated that more than half of the world’s population has internet access at home (134); the number in Sweden is 98% (135). In Sweden, more than 99% of people aged up to 55 years are internet users. Usage then decreases with increasing age, but more than 50% of citizens above 76 years of age are reported to be internet users. The internet may be an important tool used by individuals diagnosed with a disease to become informed and interact with others. A recent study suggested that individuals with cancer used the internet to a large extent when looking for cancer information, e.g., through information sites, blogs, forums, and social media (136). In addition, individuals with cancer considered the service “My healthcare contacts” to be the most valuable source of information on the internet (136). “My healthcare contacts”

26 is part of the most comprehensive internet-based health information technologies aimed at Swedish citizens (1177.se). Citizens access the site for information, to reach healthcare providers, to see their medical records, and, where applicable, to receive treatment. A new pilot-tested tool in this context, aimed at individuals with cancer, is “My Care Plan Cancer” (137). It will hopefully be implemented into standard care shortly and may be an important empowerment tool, as it aims to increase opportunities for insight, participation, and communication. “My Care Plan” may be defined as an internet-based interactive communication application connected to personal medical records. (In- ternet-based interactive communication applications are defined in the section below.)

All the technical solutions mentioned above can be included in the broad concept of e-health, involving many different types of services (including those not delivered via the internet or used solely by healthcare professionals). In 2005, a systematic review of definitions (138) found 51 different, often broad, definitions of e-health. The National Board of Health and Welfare in Sweden defines it as “to use of digital tools and the digital exchange of information to achieve and maintain health” (139). The World Health Organization has a similar definition: “the use of information and communication technologies (ICT) for health”. The national strategy for nursing efforts within e-health (140) suggests that nursing interventions should include patient participation and facilitate cooperation between professions.

Internet-based interventions promoting health in cancer Internet-based interventions targeting individuals with cancer may focus on the psychosocial, practical, physical, or psychological aspects, or a combination of these. Interactive Health Communication Applications (IHCAs) are internet-based (or computer-based) “packages” that combine health information with support in some form to promote knowledge, self-care, healthy behaviour, information exchange, and emotional support. A Cochrane review (141) on the effects of IHCAs targeting people with chronic diseases, including cancer, stated that IHCAs could lead to increased self-efficacy, positive behavioural changes, symptom burden improvements, and better health, concluding that more studies were needed for specific diagnoses, such as cancer. A review (142) on the more general concept of internet-based support programmes in cancer settings drew the conclusion that programmes including multiple services seemed to be the most effective. Symptom improvement post-intervention was mainly found regarding fatigue, social support, and distress. If an internet-based intervention is not tailored and adjusted to fit the needs of the intended target population, it will not be used; thus, the tailoring may be seen as a predictor of outcomes. Interventions that are not useful might fall under the rather harsh description “research waste” (143). According to recommendations, one way to avoid “waste” would be to include clinicians and patients

27 in research more. Other recommendations when developing these interventions include the use of best external evidence and theory (114). An example of a past internet-based intervention including the intended users in the development is WebChoice (144), using focus groups to determine intervention content (145). WebChoice is an IHCA, targeting individuals with breast or prostate cancer to help them self-manage their illness by providing, e.g., self- assessments, tailored symptom self-management support, and peer support. The outcomes showed significant within-group improvements regarding depression in the intervention group. Between groups, differences in symptom distress were significant. CHESS (146) is another example of such an intervention, targeting females, aged 60 years or less, with breast cancer. In CHESS, one of the goals was to include as many women as possible with lower levels of education and income, since they are a group known to be less likely to obtain information and participate in own care. The IHCA contained information, possibilities to ask questions, and peer support. The results from the RCT (147) indicated that the intervention increased social support and possibilities to participate in own care and thereby served women with lower levels of education and income well. However, none of these interventions targeted individuals with specific symptoms and there is also a lack of studies including clinical recruitment and stepped care models, providing individuals with support relevant to their needs.

Cognitive behavioural therapy delivered via the internet (iCBT) has been used under several conditions, with outcomes comparable to face-to-face CBT, regarding both anxiety and depression in the general population (148, 149). iCBT may be delivered as a therapy guided by a healthcare professional, or as an automated, self-guided programme. Guided iCBT seems to be the most efficient in chronic illnesses (150). In the cancer context, guided iCBT has shown positive effects regarding fatigue (151) and depression (152). The most recent RCT, iCanADAPT (153), focusing on the early stages of cancer (web- based self-referral > 6 weeks post-diagnosis with primary endpoint at three months) found positive effects regarding both anxiety and depression.

Understanding usage in internet-based interventions Non-usage and difficulties defining usage (154, 155) are well-known problems in internet-based interventions, and assessments of adherence may require multidimensional measures. Socio-demographics are sometimes used as explanations, suggesting that, e.g., those with lower education and lower socioeconomic status are less likely to make use of interventions (156). How- ever, not many interventions have focused specifically on the needs of the traditionally underserved. Thus, it seems relevant to focus on intervention parts that may be influenced to fit the intended users, e.g., the design, focus, and usability of the system. The Technology Acceptance Model (TAM) is an

28 information systems theory that is used to explain the acceptability and predictors of using new technologies (157). The theory concerns behaviour intention, i.e., factors decisive to if a person uses a system or not. According to TAM, the key factors include the perceived usefulness, perceived ease of use, and external factors. The usefulness may be defined as the degree to which a person believes that the use of a system will imply a gain for them, or, perhaps, how interested they are in the contents. Ease of use refers to how much effort is needed to use a system, i.e., how free it is from difficulties. Within TAM, it is further suggested that if the content within a system is regarded as valuable to a person, they might have a higher level of tolerance with regard to ease of use. Thus, usefulness seems to be a stronger motivator, although a balance between the two is desirable. Further, self-efficacy – as previously described within the SCT – has a close relationship to ease of use, i.e., the individual’s own perception of their abilities. The external factors are context-based, meaning that they might be found in explanatory models describing the specific sets of needs and usage predictors in the cancer context (158). Defined predictors of usage involve not only sociodemographic factors but also, e.g., psychosocial and disease- and intervention-related factors. Understanding the needs and predictors of use among the intended users may be looked upon as a foundation determining usefulness, as described within TAM.

29 Rationale for the thesis Cancer is a complex disease, with the potential of affecting all aspects of a human life. If individual care needs are not meet, the consequences may be costly, both from an individual perspective of suffering and from a societal perspective. The time surrounding the diagnosis and treatment of cancer are known to be highly stressful, leading to increased needs of support. Multiple barriers prevent individuals from getting support. These barriers include the non-detection of symptoms, the multifaceted needs of individuals, and a lack of diversity in the current resources offered. Most require physical visits to the hospital. Current routes of care delivery, in combination with the strained situation within healthcare organisations, may be contributing, and the high occurrence of unmet needs in the cancer population suggests that new methods are required.

E-health interventions have shown the potential to provide easy access to a comprehensive range of support and may be one solution to the current limited accessibility of rehabilitation resources. However, several weaknesses in previous psychosocial interventions, both internet-based and face-to-face, have been identified, such as non-evidence-based interventions aimed at those without established needs. In addition, intervention content is often poorly described and individuals with lived experience are seldom included in the development of interventions.

Thus, there is a lack of knowledge regarding how to develop clinically relevant internet-based psychosocial support, and a resulting lack of knowledge about the effects and experiences of such interventions in clinically recruited individuals diagnosed with cancer. It is crucial to bridge this knowledge gap in order to take advantage of the possibilities that the internet may offer in cancer care.

30 Overall and specific aims

The overall aim of this thesis is to explore and evaluate design, delivery, and effectiveness of internet-based stepped care for individuals with cancer and self-reported symptoms of anxiety and/or depression. The specific aims of Pa- pers I–IV were:

I To describe the development and components of the first step in a co- created internet-based stepped care intervention targeting individuals with cancer and self-reported symptoms of anxiety and depression. A further aim was to explore the acceptance of the first step among a small group of users.

II To evaluate the effects of internet-based stepped care on anxiety, depression, post-traumatic stress, and health-related quality of life in individuals with cancer and self-reported symptoms of anxiety and depression, compared with care as usual.

III To explore the user experiences of internet-based stepped care with regard to the intervention content and relevance.

IV To explore the user experiences of internet-based stepped care with regard to the delivery, design, and structure of the intervention.

31 Methods

Design This thesis is based on studies conducted within the Uppsala University Psy- chosocial Care programme (U-CARE). U-CARE is a strategic research ven- ture supported by the Swedish government with the main purpose to offer internet-based support to prevent and reduce psychosocial health problems among individuals with physical illness. All interventions are delivered via the web-based U-CARE Portal developed within the programme.

Study 1 encompassed a co-creation development process resulting in the interactive support provided as the first step in an internet-based stepped care intervention (iCAN-DO). The effects of iCAN-DO were investigated in a randomised controlled trial; AdultCan (clinicaltrials.gov NCT-01620681) targeting individuals newly diagnosed with breast, colorectal, or prostate cancer and with self-reported symptoms of anxiety and depression. Study 1 made up the material for Papers I-II. In Study 2, aspects of usefulness, relevance, and usability in iCAN-DO were explored through qualitative interviews which made up the material for Papers III-IV. An overview of the specific design, methods, samples, and analyses is shown in Table 1.

The U-CARE Portal U-CARE projects deliver interventions using the U-CARE Portal and is not connected to personal medical records. Study participants and researchers log in to the portal either through a two-step verification (additional code sent by SMS) or using a mobile BankID. A helpdesk for technical support is available for participants in the ongoing studies. In addition to delivery of interventions, the portal also allows administration of PROMs. Currently, the system em- ploys the Bootstrap framework, a responsive design adapting each page to the screen size of a user’s device (i.e., tablet, smartphone, computer). To enable efficient use of the programmes developed within U-CARE (post-intervention) the intervention materials in the portal are licensed under Creative Com- mons.

32 Table 1. The design, data collection methods, samples, and analyses in Papers I–IV.

Papers Design Data collection Sample Analyses

I. Theoretical/descriptive Log data from the U- The 39 first participants Descriptive CARE Portal in AdultCan IG

II. RCT PROMs, log data, 245 participants in Descriptive register AdultCan IG and SCG Linear mixed models Multiple linear regression Randomisation tests

III. Qualitative, descriptive Semi-structured 15 participants in Content analysis interviews AdultCan IG

IV. Qualitative, descriptive Semi-structured 15 participants in Content analysis interviews AdultCan IG Abbreviations: RCT, randomised controlled trial; PROMs, patient-reported outcome measures; IG, intervention group; SCG, standard care group.

Participants and settings Study 1 (Papers I–II) Individuals were recruited to AdultCan from four hospitals in mid-Sweden between the years 2013 and 2016. A research nurse/assistant identified and approached eligible individuals consecutively when they were at clinical ap- pointments in hospital. Eligible individuals were newly diagnosed (< 6 months prior) adults (> 18 years) with breast, colorectal or prostate cancer, or a relapse in colorectal cancer. Criteria for exclusion were the inability to speak and read Swedish, a low Karnofsky performance status (< 40) (159), short expected survival (< 3 months), cognitive impairment, and an ongoing severe depression or risk of suicide according to the MADRS-S (160) (described below).

If they chose to participate, an informed consent form was signed by the participant, who later received further instructions and login details through automated e-mail. At their first login to the U-CARE Portal (described below), they were prompted to complete the screening instrument HADS (52). Partic- ipants with symptoms of anxiety and depression or both (> 7 p on any of the HADS subscales) were then randomly assigned within the U-CARE portal using a permuted block method. This means that the randomisation process was completely computer-generated and concealed from all within the research group. Further, randomisation was stratified for curative or palliative treatment. The screening process was completed for 909 individuals, of which 664 (73%) scored below the cut-off on both HADS subscales (< 7). These

33 individuals were allocated to a longitudinal study, not included in this thesis. Upon randomisation to standard care (SC) or intervention group (iCAN-DO), the MADRS-S was completed to detect severe cases of depression or risk of suicide. Ten participants reported a high level of severity (generating an alert in the portal) and were contacted by a clinical psychologist for assessment. However, none of the individuals contacted was assessed as needing referral to psychiatric care and all could therefore remain in the study. In total, 245 participants were allocated to the RCT. The first 39 participants randomised to the iCAN-DO group constituted the sample in a first acceptance test, described in Paper I. Figure 3 shows the allocation of participants, including assessment points and loss to follow-up.

Computer literacy was not an exclusion criterion, but naturally determined if an individual could participate in the study. Eligible individuals were informed that they needed computer/internet access in order to participate and declined to participate if they were lacking this. Those who declined participation were asked if they would voluntarily fill out a response analysis and 950 individuals consented to do this.

Assessed for eligibility n = 3,273

Declined to participate n = 1,188 Did not meet inclusion criteria n= 662 Not approached due to administrative or logistic failure n = 140

Provided informed consent n = 1,283 Completed baseline n = 909

Randomised n = 245 (> 7 points on any of the HADS subscales)

iCAN-DO n = 124 Standard care n = 121

Breast cancer n = 85 Breast cancer n = 71 Prostate cancer n = 24 Prostate cancer n = 28 Colorectal cancer n = 15 Colorectal cancer n = 22

Completed assessment n = 99 T1 Completed assessment n = 100 Withdrew from trial = 0 Withdrew from trial = 2 Did not complete questionnaires *= 22 One month Did not complete questionnaires *= 19 Deceased = 3 Deceased = 0

Completed assessment n = 78 T2 Completed assessment n = 87 Withdrew from trial = 4 Withdrew from trial = 0 Did not complete questionnaires *= 35 Four months Did not complete questionnaires *= 31 Deceased = 1 Deceased = 1

Completed assessment n = 65 T3 Completed assessment n = 75 Withdrew from trial = 0 Withdrew from trial = 0 Did not complete questionnaires*= 51 Seven months Did not complete questionnaires *= 43 Deceased = 0 Deceased = 0

Completed assessment n = 61 T4 Completed assessment n = 80 Withdrew from trial = 0 Withdrew from trial = 0 Did not complete questionnaires*= 54 Ten months Did not complete questionnaires *= 37 Deceased = 1 Deceased = 1

Figure 3. Flowchart of enrolment, participants, and missing assessments in U-CARE AdultCan. * Number of participants who did not complete the questionnaires at this timepoint without having actively withdrawn. Abbreviations: HADS=The Hospital Anxiety and Depression scale.

35 The intervention The stepped care intervention (iCAN-DO) comprised two steps, available for participants in the intervention group. Nurse-led interactive support (Step 1) was available immediately after randomisation and throughout the study period of 10 months. A ten-week iCBT programme (Step 2) was offered at repeated occasions, coinciding with the self-reporting of symptoms of anxiety and depression. Thus, participants were only offered iCBT if they had results on a HADS subscale (52) that indicated remaining symptoms (> 7). The time- line is shown in Figure 4.

Step 1, available throughout the 10-month study period T1 T2 T3 T4 Baseline 1 month 4 months 7 months 10 months Study end

Step 2, iCBT Step 2, iCBT Step 2, iCBT offered if T1 offered if T2 offered if T3 = = = HADS > 7 HADS > 7 HADS > 7 Figure 4. Overview of data collection points and access to steps 1 and 2.

Development and testing of the first step in iCAN-DO The development of the first step in iCAN-DO had the working name NILS and was eventually named “interactive support” (IAS). The development process built on an evidence-based triad (161), including the use of best scientific evidence, clinical expertise, and input from presumptive users, in this case represented by individuals with lived experience of cancer, recruited from patient associations. This was done to achieve an intervention with high relevance to its users, and good compliance with standard healthcare, thereby reducing the risk of “research waste”. During the development of Step 1, a multi-professional group consisting of clinically active professionals was formed. The group assisted in shaping some of the materials. In addition, individuals with lived experience of cancer (the same diagnoses as the intervention was directed at) assisted the research group monthly during a period of one year. They reviewed all intervention contents with regard to various aspects and tested parts of the U-CARE Portal, such as the forums. Further, content creation was guided by propositions of existing theories, mainly Orem’s S-CDNT (118) and Bandura’s SCT (162), with elements of psychoeducation.

36 Interactive support, Step 1 The low-intensity support was designed to provide a relevant basis for Step 2 (described below) and comprised various parts to attract participants with differing preferences. The most prominent feature, the library, contained separate lectures on each diagnosis (breast, colorectal, and prostate cancer), as well as 16 modules (Table 2), including in-depth psychoeducational content regarding diagnosis, treatment, and common symptoms/impairments/problems, all with information on self-care workarounds. The contents in the library were aimed at explaining and normalising problems and providing guidance on how to deal with them, with the goal of increasing the individual’s general knowledge, commitment, and abilities. Further, participants had the possibility to “ask an expert” (a nurse with experience in oncology care). Since the portal was not connected to medical records, participants were informed that no specific advice – for instance regarding prescribed drugs – could be given. Answers were delivered straight to the participant who asked the questions via an internal messaging function in the portal within 2 working days, giving the nurse a possibility to consult other professions (e.g., dietician, physician) if needed. Questions and answers of general interest could also be anonymised and published in a frequently asked questions (FAQ) section, for other participants to read. In addition, a moderated peer support section allowed participants to discuss with each other in forums or chat. A nurse moderated the portal daily, looking for problem-oriented discussions as well as checking logs for predefined words that could imply suicidal thoughts or severe depression. In addition to having diversity regarding content, the materials were available in different media types, such as written texts, audio-visual presentations, vid- eos, and slideshows. Nurses were presented in brief with name and photo.

Internet-delivered cognitive behaviour therapy, Step 2 The second step was a ten-week iCBT programme guided by a psychologist. Treatment comprised 15 modules (Table 2), with areas relevant for individuals with cancer, such as fatigue, worrying, and depressed mood. After completing an introductory module, participants could choose which modules to work with during the ten weeks. Each module involved “homework assignments”, e.g., self-monitoring, that participants worked with and sent to the psychologist. Feedback was delivered weekly, and the psychologist was also available for questions. As in Step 1, all communications took place through written messages and the psychologist was presented with name and photo.

Table 2. The contents and features of the internet-based stepped care in iCAN-DO, aimed at individuals with cancer and concurrent symptoms of anxiety and depression. Step 1 Step 2 Interactive support (IAS) guided by a Internet-delivered cognitive behaviour nurse; all communication was in the form therapy (iCBT) guided by a psychologist; of written messages via the all communication was in the form of U-CARE Portal. written messages via the U-CARE Portal. Who and when? Available for all participants in the inter- Offered to participants with persistent vention group after randomisation. symptoms of anxiety and/or depression Duration 24 months. (> 7 on either HADS subscale) after using Step 1. Duration 10 weeks. Contents The main feature of Step 1 was a library In Step 2 of the intervention, a ten-week containing psychoeducational information iCBT treatment programme was provided. and self-care strategies in 16 modules. In The treatment contained 15 modules that addition to the library, there was a peer comprised written texts, audio-visual support section and “Ask an expert”/FAQ presentations, and video clips. section. The Library modules: The iCBT modules:  Radiotherapy  Introduction to iCBT  Chemotherapy  Depressed mood  Endocrine treatment  Tension  Nausea  Stress  Loss of appetite  Negative thoughts  Cancer-related fatigue  Fatigue  Cancer-related pain  Relations  Reactions to a crisis  Pain  Anxiety  Negative body image  Depressed mood  Fear  Sleep disturbances  Sleep disturbances  Sexual problems  Sexual problems  Healthy eating habits  Worry and rumination  Physical activity  Problem-solving  The social insurance system  Final/summarising module  Tobacco use Abbreviations: IAS= interactive support (Step 1), iCBT=internet-delivered cognitive behaviour therapy (Step 2), HADS= the Hospital Anxiety and Depression scale.

Definition of standard care Both the SC group and iCAN-DO group had access to standard care. As stated in the introduction, this include the regular support from encounters with healthcare staff, the possibility of prescription drugs targeting, e.g., depression, and the possibility of psychosocial support by self-referral to a counsellor or the hospital church.

Data collection As stated above, all PROMs were administered via the portal and participants received e-mail notifications at the start of an observation point. In addition to

38 the validated PROMs used, participants answered questions regarding socio- demographics and perceived benefits of using Step 1 in questionnaires specifically designed for AdultCan. The perceived benefits questionnaire contained questions regarding, e.g., level of knowledge and use of self-care strategies and was completed at one month (after using Step 1, before iCBT was offered). All data regarding participants’ diseases and treatments were obtained from national quality registers. Table 3 shows an overview of the timepoints and validated PROMs used.

Table 3. Overview of variables, instruments, and timepoints in AdultCan. Variable Instruments T0 T1 T2 T3 T4 Baseline 1 month 4 months 7 months 10 months

Anxiety The Hospital Anxiety and Depression x x x x x Scale- Anxiety subscale (HADS-A) The Spielberger State-Trait Anxiety Inven- x x tory State subscale (STAI-S) Depression The Hospital Anxiety and Depression x x x x x scale- Depression subscale (HADS-D) The Montgomery Åsberg Depression Rat- x x ing Scale (MADRS-S) Health-related European Organisation for Research and x x x x x quality of life Treatment of Cancer Quality of Life Ques- tionnaire Core 30 (EORTC QLQ-C30) Insomnia The Insomnia Severity Index (ISI) x x x x x Fatigue The Functional Assessment of Chronic Ill- x x x x x ness Therapy Fatigue Scale (FACIT-F) Post-traumatic The Post-traumatic Stress Disorder Check- x x stress list-Civilian version (PCL-C) Environmental The Environmental Reward Observation x x reward Scale (EROS)

Main outcome measures

The Hospital Anxiety and Depression Scale (HADS) The main outcomes were anxiety and depression as assessed with the HADS (52). The HADS contains 14 statements/items and may be interpreted as a total (HADS-T), with distress as the outcome, or using the subscales HADS- A (anxiety) and HADS-D (depression). The HADS-T scores range between 0 and 42 points, with 0–14 points categorised as non-distress, while 15–42 points suggest distress (53). As for the anxiety and depression subscales, scores range between 0 and 21 for each. Recommended guidelines for interpretation of subscale scores are that values < 7 points do not indicate symptoms, values of 8–10 points indicate possible symptoms and values of > 11 points indicate likely symptoms. Individuals within these levels may be classified as non-cases, doubtful cases, and clinical cases. The items are focused on the preceding week and agreement with the statements are rated on a Likert scale of 0–3. According to a systematic review (163), the sensitivity and specificity for the HADS-A and -D are 70% and 83%, respectively, when using < 7 points as the cut-off. The MCIDs for the HADS-A and -D have been suggested to be 1.32 points and 1.40 points, respectively (164). The HADS is recommended within PROMS-Cancer Core (73).

39 The Montgomery-Åsberg Depression Rating Scale (MADRS-S) In addition, depression was assessed with the self-reported version of the Montgomery-Åsberg Depression Rating Scale MADRS-S (160), originally designed to detect change in primary depressive illness in accordance with the Diagnostic and Statistical Manual of Mental Disorders (DSM-V). The MADRS-S consists of nine items on a seven-point Likert scale, with total score ranging from 0 to 60 points. One of the items (number nine) may be used as a specific indicator for the risk of suicide. Recommended guidelines for interpretation are that 13–19 points indicate mild depression, 20–34 points moderate depression and > 34 points indicate severe depression. The recall period focuses on the preceding three days. The MADRS-S has been used in cancer populations, but evidence of its validity there is lacking. However, it has been suggested that the MADRS correlates with findings of hopelessness/helplessness (closely related to depression) in females with breast cancer (165).

The Spielberger State-Trait Anxiety Inventory State subscale (STAI-S) Anxiety was also assessed with the STAI-S (92) (state; anxiety at the moment) consisting of 20 statements to be rated on a four-point Likert scale. The time of recall for the STAI-S is “at the moment”. Higher scores indicate higher levels of anxiety. The suggested cut-off point to detect clinically relevant symptoms is 39–40 (166). The STAI-S has previously been shown to be valid in individuals with chronic illnesses (92), with the highest sensitivity and specificity at cut-off levels of 54/55 points. The STAI-S is recommended within PROMS-Cancer Core (73).

The Environmental Reward Observation Scale (EROS) In addition to instruments assessing the severity of depression symptoms, the risk of worsening is a pertinent measure. In the behavioural approach to depression, behaviour activation is an important component. An increase of the individual’s positive exchanges with the environment is seen as essential in the improvement of depressive symptoms. Therefore, the Environmental Re- ward Observation Scale (EROS) (167) was used to assess the risk of worsening in depression as measured through levels of environmental reward. It is a short measure, consisting of 10 statements, e.g., “In general, I am very satisfied with the way I spend my time”. Statements are rated on a four-point Likert scale and scores range from 20 to 80; a high score indicates high environmental reward.

40 Secondary outcome measures European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) The EORTC QLQ-C30 (version 3) (82) was used as a measure of HRQoL. The questionnaire contains 30 questions focusing on the preceding week. Questions regarding symptoms and functions are measured on four-point Lik- ert scales (0 = not at all, 4 = very much). The last two items, forming the overall global health status, are measured on seven-point Likert scales (1 = very poor, 7 = excellent). Using the EORTC scoring manual (168) and a program for statistical analysis, the 30 questions are merged into five scales of functioning (physical, role, emotional, social, and cognitive) and three scales regarding symptoms (fatigue, pain, and nausea/vomiting), while six items stay at a single-item level (dyspnoea, insomnia, constipation, diarrhoea, loss of appetite, and financial impact). Further, the responses to items and subscales are transformed into a total scale of 0–100 to ease statistical analysis. In the manual, several publications are also recommended by the EORTC for interpretation of results, e.g., MCIDs (169). The QLQ-C30 is recommended within PROMS-Cancer Core (73).

The Insomnia Severity Index (ISI) Insomnia was assessed using the Insomnia Severity Index (ISI) (87), containing seven statements regarding, e.g., sleep onset, duration, and feelings of being rested. The recall period focus on the preceding two weeks and are given on a five-point Likert scale. Guidelines for interpretation are that values between 8 and 14 suggest mild insomnia, values of 15–21 suggest insomnia, and values of 22–28 suggest severe insomnia. A previous study (88) on individuals 1–2 years post-cancer revealed a mean score at 16.69 (SD 4.47). The ISI is recommended within PROMS-Cancer Core [69].

The Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) The Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) (89) was used to assess CRF. This is a short tool that may be used as an extension to the FACT-F or as a stand-alone instrument. The FACIT-F contains 13 items measured on a four-point Likert scale. The recall period covers the preceding seven days, and the score ranges from 0 to 52, with a high score representing less problems. Guidelines for MCIDs are available (170). The FACIT-F is recommended within PROMS-Cancer Core (73).

The Post-Traumatic Stress Disorder Checklist-Civilian version (PCL-C) For assessment of post-traumatic stress, the Post-Traumatic Stress Disorder Checklist-Civilian version (PCL-C) (94)was used. It consists of 17 items rated on a five-point Likert scale. The PCL-C has since been updated and the present

41 version (PCL-5) consists of 20 items. Suggested cut-offs differ between populations, e.g., between rape victims, soldiers, and civilians. The cut-off for the latter group is suggested to be 44 points.

Determination of sample size In a previous study (38) using the HADS as a screening tool, the mean values for individuals with cancer scoring above the cut-off level (> 7) were 10 p for the HADS-A and 8 p for the HADS-D. Meanwhile, the MCID for the HADS in a similar population is considered to be a change/improvement of 20% for subscales (164). To detect a 20% mean score difference (HADS-A 2 points, HADS-D 1.6 points) between groups with 80% power at a significance level of 0.05, 65 participants were needed per arm. Thus, the risk of type 2 errors was at a 20% level and that of type 1 errors was at a 5% level. Since internet- based studies are known to have high drop-out rates, twice this number was recruited.

Statistical analyses When data had been extracted from the U-CARE Portal, all scores were pro- cessed based on the instructions for each respective instrument (e.g., the EORTC scoring manual). The questionnaires completed at baseline and after ten months (MADRS-S, STAI-S, PCL-C, and EROS) were analysed with multiple linear regression. The questionnaires completed at all five timepoints (HADS, EORTC QLQ-C30, ISI, and FACIT-F) were analysed with linear mixed models (LMM) fitted using restricted maximum likelihood. This was done to reduce the risk of bias in the model. The LMM included all timepoints in the model. Both models were adjusted for baseline values and level of education. In addition to these analyses, data on categorical level (HADS classifications) were analysed with randomisation tests (171), including multiple resampling. All analyses were performed in R version 3.4.2 and SPSS version 25. In addition to the statistical analysis, defined MCIDs (78) were used to assess the clinically significant differences and a responder analysis (catego- rising results as individuals improved, not changed, or deteriorated) was conducted in individuals with complete data on the HADS.

Analysis populations All outcomes were analysed according to intention-to-treat (ITT), including participants with complete baseline values. The method of imputation was the last rank carried forward method (LRCF). This allows for missing participants to keep their rank in relation to the others, which the last observation carried forward method does not. The risk for false high or low imputations is thereby reduced. In addition to analysis based on ITT, analyses of the primary outcome (HADS) were carried out in the population “modified ITT”, including participants with a baseline value and at least one more follow-up value, again using LRCF for the remaining timepoints. Further, “complete cases” (those with all

42 assessments complete) were analysed separately. For the responder analysis, a fourth population was studied: all individuals with complete data on the HADS at both baseline and ten months. The purpose of using the additional populations was to perform a sensitivity analysis, i.e., to see if results were robust in populations other than ITT.

Study 2 (Papers III–IV) Participants and recruitment The participants in Study 2 were recruited from the AdultCan RCT. As data from the RCT showed that most participants used Step 1 (IAS) in iCAN-DO (105/124, median 9 times), and few of those who were offered Step 2 (iCBT) accepted (7/82), a purposeful selection was performed. The main purpose of the selection was to include individuals who had used both Step 1 and Step 2, and also to include individuals who had declined to use Step 2. Logs from the U-CARE Portal were studied to perform this selection. The number of participants was not predefined, but instead determined during the performance of the interviews. Eligible participants were sent a letter with information about the study and that the interviewer would call them within a few days to give further information and to ask if they were willing to participate. They were also informed that they could phone or e-mail the principal investigator if they did not want to get this phone call (no one did so). If the informant consented to be interviewed, a time and place for this was determined during the phone call. The place where the interview would be conducted was chosen by the participant. In total, 20 individuals were approached, two declined to participate (one due to lack of time, and one due to not wanting to talk or think about the cancer anymore). Three individuals could not be reached. All in all, 15 individuals were interviewed. They had all had access to iCAN-DO for at least seven months (the last offer to participate in Step 2 was at seven months). The total sample in the AdultCan RCT was dominated by females with breast cancer, which is also reflected in the sample of Study 2. However, the sample for Study 2 differed from the total sample regarding age, level of education, and general online activity. Further, two of the participants described themselves as very inexperienced computer users. Participant characteristics are shown in Table 4.

Data collection In addition to retrieving information from logs in the U-CARE Portal regarding levels of activity during the selection process, information regarding the participants’ characteristics was also retrieved from self-report questionnaires in the portal. At the time of the interviews, participants were asked to describe their computer literacy and their general online activities.

43 Table 4. Participants’ characteristics, n = 15.

Age in years, mean (min–max) 59 (37–69) Gender Female 10 (breast cancer) Male 5 (1 colorectal cancer, 4 prostate cancer) Relationship status Married or partner/living with someone 12 Married or partner/living alone 1 Widowed 1 Single 1

Level of education Elementary or middle school 3 High school 1 University up to 3 years 6 University more than 3 years 5

Working situation Working 10 Retired 4 Early retirement 1

General online activity outside iCAN-DO Daily, no social media 4 Daily, active on social media, lurking 7 Daily, active on social media, participating 4

Activity in Step 1 (IAS) Opening material (all sections) a > 20 times 6 Opening material (some sections) > 20 times 7 Opening material (some sections) < 10 times 2

Step 2 (iCBT) No 9 Yes 6

NOTE: a. Library, peer support, FAQ/ask an expert. Abbreviations: IAS = interactive support, iCBT = internet-delivered cognitive behavioural therapy.

44 Interviews The interviews were conducted by the first author, a registered specialist nurse with some experience in oncology nursing. The interviewer had been involved in parts of the development of iCAN-DO and was also assigned to moderating forums and answering questions in the “Ask an Expert” section within Step 1. (One of the participants had had a question answered by the interviewer.) At the start of interviews, the interviewer assured the participants that the purpose of the interviews was to get a better understanding of how the contents and delivery was perceived, and that she was interested in both pros and cons. If the participant wanted, they could have iCAN-DO open in front of them during the interview. All interviews were tape-recorded and an interview guide with open-ended questions was used. The interview guide was constructed to help the interviewer achieve stability, e.g., to cover the same topics in all interviews. The interview guide was divided into two parts, one regarding the more technical usability issues and one regarding the contents and relevance. All questions were open-ended, e.g., “Can you tell me about your experiences of using the content within iCAN-DO?” All parts of iCAN-DO were covered in the guide. Further, follow-up questions were based on the responses and were used to get further, deeper descriptions. The follow-up questions could be probing, e.g., “How did that make you feel?” or specifying, e.g., “Okay, and how did that help you?” When necessary, interpreting questions were used, e.g., “Do you mean that this fact made you use that function more often?”

Data analysis The first author transcribed all interviews verbatim, and inductive content analysis (172) was used to analyse the manifest content of the interviews. Transcription was conducted in close connection to each interview. This was done to have a clear memory of the situation, but also as a learning experience (e.g., to determine when it would had been better to stay quiet or when to penetrate a subject further). In the next step, each text was read repeatedly and divided into meaning units. The meaning units were then condensed, i.e., un- necessary words were removed while the original core was preserved. After condensation, each meaning unit received a code describing its key message. The codes that appeared to be related to similar contents were then assigned into subcategories, categories and in one case a theme, describing the contents at a higher abstraction level.

Ethical considerations (Studies 1 and 2) In all scientific studies involving humans, measures must be taken to ensure the protection of their rights. Through history, this has not always been self- evident. Perhaps the most famous example of ethical misconduct is that of the sinister medical experiments resulting in the Nuremberg Code (173), one of

45 the first attempts to form an international ethical code for research. The Nu- remberg Code contains ten ethical principles regarding experiments involving humans, which may seem very basic and obvious today (e.g., protecting the experimental subject from disability or death), but it had a great impact and led to several, more modern guidelines. One of those is the Declaration of Helsinki, an international standard approved in 1964 by the World Medical Association (174). The declaration states that the wellbeing and health of the participants must always be the main consideration in studies. Both of the studies in this thesis were conducted in accordance with the ethical standards of the declaration. In Studies 1 and 2, the possible burdens to participants were thoroughly discussed in both planning, execution, and reporting of studies. The work described in Paper I may be seen as part of these considerations, providing the participants with an intervention based on scientific principles and best knowledge, which is one of the principles within the Declaration of Helsinki. The literature did not suggest that any content in iCAN-DO could involve potential, serious harm to participants. Some issues, such as the risk of feeling worse when participating in online discussion forums (147), were managed by moderating forums in iCAN-DO. Another risk evaluated was that of including individuals with severe depression or risk of suicide, when iCAN- DO might not provide them with the appropriate support. Therefore, the MADRS-S was added as a way to detect and refer those individuals to spe- cialised psychiatric care. Upon choosing to participate in the studies, participants provided written informed consent, an act that has its foundation in the Nuremberg Code, where the first principle is that of participate of own free will. All participants were provided with a copy of their informed consent along with written information concerning the study and their rights. Partici- pants in both studies had the option to withdraw from the studies at any time.

Other risks that were discussed were those of feeling worse when completing PROMs and disappointment in not getting access to the support provided within iCAN-DO, either by being randomised to standard care or by lacking computer literacy. During enrolment, participants who could not participate due to lacking computer literacy could be referred to a regular counsellor by the research nurse/assistant. As stated earlier, all participants had access to standard care services and were free to withdraw at any time. Further, the U- CARE Portal is located behind Uppsala University’s firewall and the personal identity numbers of participants were stored in a server separated from the intervention server. Thus, the researcher did not have access to this information without requesting a “security breach”, which could be done in the cases of identified severe depression or suicide risk.

In Study 2, no anonymity could be offered, since interviews were conducted face-to-face, often in the participant’s home. What could be offered was con- fidentiality, e.g., the interviewer presented herself with name or waited for the participant to introduce her in case of encounters with any other persons in

46 connection to the interviews. Participants were treated with friendly respect and the interviewer strived to keep an open mind during interviews, not judg- ing participants based on what was said or implied.

A tape recorder was used during interviews and was kept locked up in a cabi- net approved for research data. In addition, no personal data such as name or personal identification number were recorded. The transcripts contained only an anonymised number for each participant. Both studies were approved by the Swedish Ethics Review Authority (Study 1; no. 2012/003, Study 2; no. 2012/003/9).

47 Results

The development (Paper I) The results from Paper I are a description of the rationale behind the evidence- based triad, and how the work was conducted. They also include a description of the first, lower intensity support (IAS; Step 1) and the theoretical motives behind it. This approach will hopefully serve to expand on the earlier lacking descriptions of interventions. During development, the evidence-based triad was applied in the groups of clinically active professionals, representatives with lived experience, and the research group (Figure 5). The S-CDNT guided the educative, supportive nursing approach, and self-care deficits were seen as naturally occurring when a patient was newly diagnosed with cancer. All the materials involved psychoeducation, a method used within iCBT to nor- malise and explain symptoms and feelings.

Best external evidence Individual clinical expertise Patient values and expectations

The resources within the The resources within the group The resources within the lived research group of clinically active professionals experience group

Figure 5. The application of the evidence-based triad.

To provide participants with a new self-care agency, various materials in both didactic and narrative forms (Figure 6) were provided in a Library. The need for different forms of delivery of the content was highlighted by the lived experience group, who also stressed the need for simplicity and contents with clear connection to cancer. The materials in the library consisted of three diagnosis-specific lectures held by physicians, and 16 modules following the typical course of events in cancer. The subjects of the modules were identified within the evidence-based triad, e.g., nausea, sleep disturbances, and anxiety (all modules are presented in the methods section of this thesis). Each module started with a main lecture with a psycho-educative approach, providing explanation of the problem/symptom in question, followed by basic measures of self-care, providing the user with self-care agency. The main lectures were

48 followed by specific self-care elaborations based on the module content. Thus, the module on depression was followed by self-care measures involving, e.g., behavioural activation, while the module on chemotherapy was followed by self-care measures concerning, e.g., mucositis, nausea, and xerostomia.

Figure 6. Examples of content in the Library (Step 1).

If a participant needed further information or additional support to perform self-care, they could pose own questions as well as read the anonymised questions of others (in the FAQ) within the function “Ask an Expert”. The new knowledge could potentially lead to increased confidence, or self-efficacy, as described within the SCT. Further, constructs within the SCT suggested that learning with others, i.e., social learning (to reflect and imitate others), might enhance and reinforce behaviours. To provide such an environment, a Peer support section was created, allowing participants to interact. Since previous research has shown that unmoderated forums may cause increased distress in individuals with cancer, a nurse moderated the portal daily, looking for problem-orientated discussions. Psychoeducation was also used in the communication with participants, e.g., when answering participant questions, the answer included validation and normalising of the situation/symptom, as well as an explanation of and advice on self-care. An example of this structure is presented below:

49 “Dear X, this is a highly relevant question and it is fully understandable that you worry about it [validation]. It is very good that you already know that these symptoms may occur and want to know more about how you can prepare yourself. This may help you manage the situation better if you experience symptoms. The symptoms that you describe are common [normalising] during radiotherapy, but not everyone is affected… [statistics]… the symptoms depend on… [explanation based on existing evidence- based knowledge]. I suggest that you… [guidance to self-care or suggestion to contact regular healthcare]. I hope this answer provided you with some guidance and I wish you all the best during your treatment. Please get back to me if you want me to clarify anything or if something else comes to mind.”

Initial test of acceptance When the first 39 participants (breast, n = 29, colorectal, n = 6, and prostate cancer, n = 4) had been randomised to iCAN-DO, a decision was needed on whether to continue enrolment or to stop the study to make changes to the intervention. Activity logs were extracted from the U-CARE Portal to investigate if the participants used the first step in iCAN-DO, i.e., if the intervention was implemented as intended. Results showed that 82% had accessed Step 1 (median = 8, range 1–126) and 62% had accessed material more than twice. These numbers were compared with the numbers reported in similar studies (WebChoice (144) and CHESS (147)) and since the use of Step 1 seemed comparable to these, the AdultCan RCT was deemed fit to continue.

The AdultCan randomised controlled trial (Paper II) The descriptive data showed that non-participants (n = 1,188) were older than participants (mean: 68 vs. 61 years). The most common diagnosis among non- participants was colorectal cancer, i.e., they more often had advanced disease. The voluntary response analysis was completed by 950/1,188 and showed that 55% declined due to lack of computer/internet access or lack of computer literacy. Of the 909 participants that completed baseline 27% (n=245) had a score above seven in any of the HADS-subscales. Of the 245 participants randomised in the RCT (iCAN-DO n = 124; SC n = 121), most (69%) were included within four months of diagnosis, i.e., most received medical treatment during the study period. Regarding the sociodemographic characteristics, no statistically significant differences were seen between groups at baseline. However, at T4 (10 months), those remaining in iCAN-DO had a higher level of education than in SC, i.e., a larger proportion with lower level of education were non-responders at T4. Participant characteristics are shown in Table 5.

50 Table 5. Participants’ characteristics at baseline and ten months in AdultCan. Randomised Completed 10-month assessment Characteristic (n = 245) (n = 141)

iCAN-DO Standard care iCAN-DO Standard care (n = 124) (n = 121) (n = 61) (n = 80) Age, years Mean (SD) 57.3 (10.6) 57.0 (10.7) 58.7 (9.9) 58.4 (10.4) Sex n (%) Female 93 (75) 81 (67) 40 (66) 50 (63) Male 31 (25) 40 (33) 21 (34) 30 (37) Marital status n (%) Single/divorced/widowed 20 (16) 13 (11) 5 (8) 9 (11) Married/cohabiting 95 (77) 98 (81) 53 (87) 64 (80) Living apart together 6 (5) 6 (5) 2 (3) 3 (4) Other 3 (2) 4 (2) 1 (2) 4 (5) Education* n (%) Elementary school 18 (15) 23 (19) 8 (13) 16 (20) High school 36 (29) 42 (35) 14 (23) 28 (35) University/university college ≤ 3 years 29 (23) 28 (23) 14 (23) 20 (25) University/university college > 3 years 41 (33) 28 (23) 25 (41) 16 (20) Tumour origin n (%) Breast cancer 84 (68) 71 (59) 38 (62) 46 (57) Prostate cancer 24 (20) 28 (23) 16 (26) 22 (28) Colorectal cancer 14 (12) 22 (18) 7 (12) 12 (15) Monthly income n (%) Low < €1,244 5 (4) 5 (4) 2 (3) 3 (4) Average €1,244–4,976 82 (66) 79 (65) 36 (59) 51 (64) High > €4,976 5 (4) 3 (3) 5 (8) 2 (2) Unknown (not responded) 32 (26) 34 (28) 18 (30) 24 (30) Psychosocial support prior to inclusion n (%) Yes 7 (6) 7 (6) 5 (8) 3 (4) No 117 (94) 114 (94) 56 (92) 77 (96) Currently using psychotropic drugs n (%) Yes 24 (19) 24 (20) 14 (23) 15 (19) No 100 (81) 97 (80) 47 (77) 65 (81) Computer experience n (%) Very inexperienced 5 (4) 6 (5) 2 (3) 2 (2) Quite inexperienced 12 (10) 12 (10) 4 (6) 10 (13) Quite experienced 54 (44) 45 (38) 26 (43) 32 (41) Very experienced 52 (42) 57 (48) 29 (48) 35 (44) NOTE: Where numbers in a category do not add up to n or 100%, there is missing data. *For completed participation, iCAN-DO and SC differed regarding education (Pearson chi-square value 7.952, df 3, asymptotic significance 0.047).

51 Utilisation Of the 124 participants randomised to iCAN-DO, 105 (85%) utilised the material in Step 1 (IAS) (median visits per participant = 9, range 1–92). The Li- brary was the most visited feature (87% of 105), with most views in modules concerning radiotherapy and fatigue. The modules concerning anxiety and depression were visited by 13% and 19%, respectively (all visits shown in Figure 7). The Peer support section was visited by 65% and “Ask an expert” including the FAQ was visited by 58%. Most visits (78%) took place within four months from randomisation. The second step (iCBT) was offered to 82 of 124 (66%), 16 accepted and 7 completed at least one module.

Figure 7. Diagram showing number of participants visiting the different modules in the library (Step 1). The diagram is sorted clockwise, with the fewest visits at twelve o’clock (Tobacco use) and the most at eleven o’clock (Radiation therapy). Each ring in the diagram equals five visits. The module “Facts about your diagnosis” had diagnosis-specific contents (for breast, colorectal, or prostate cancer).

52 Perceived benefits At T1 (one month), before the first offer of iCBT, participants in the iCAN- DO group completed two project-specific questionnaires regarding perceived relevance, knowledge, and satisfaction after using Step 1. Of the 124 participants, 77% (n = 95) completed this questionnaire. Of these, 90% stated that they used the recommended self-care strategies and 76% that they perceived an increase in their own knowledge concerning the cancer and its symptoms.

Anxiety and depression Of the 909 individuals who completed baseline, 245 (27%) scored above seven on any of the HADS-subscales. Forty-four percent in the iCAN-DO group reported symptoms of both anxiety and depression at baseline, 34% of anxiety only, and 22% of depression only. Corresponding proportions in the SC group were 39%, 49%, and 12%. The ITT analysis showed a mean decrease in depression measured with the HADS-D in the iCAN-DO group that was statistically significant (LMM; p = 0.049), but not clinically significant, compared with the SC-group. The statistically significant differences regarding depression were not seen between groups in the sensitivity analysis (LMM; modified ITT p = 0.072, complete cases p = 0.495). However, in all analysis populations, a clinically meaningful decrease for the HADS-D was seen within the iCAN-DO group, but not within the SC-group (Table 6). As regards anxiety measured with the HADS-A, no statistically or clinically significant differences between or within groups were found.

Table 6. Descriptive statistics and changes over time for the HADS Anxiety and De- pression subscales for individuals randomised to iCAN-DO or standard care.

Decrease Between within group Populations Allocation Baseline 1 month 4 months 7 months 10 months group differences T0 T1 T2 T3 T4 BL–10 10 months months

ITT1 Depression Mean (SD) n = 245 Standard 7.9 (3.7) 7.7 (4.1) 7.0 (4.2) 7.1 (4.7) 7.3 (4.4) 8% care n = 121 15% iCAN-DO 8.0 (3.1) 7.3 (3.6) 6.8 (3.9) 6.5 (3.8) 6.2 (3.6) 23%* n = 124

Anxiety Mean (SD) Standard 10.2 (3.1) 9.2 (4.1) 8.9 (4.1) 9.5 (4.3) 9.5 (4.1) 7% care n = 121 4% iCAN-DO 9.2 (3.0) 7.7 (3.7) 7.9 (3.9) 8.1 (3.6) 8.2 (3.9) 11% n = 124

ITT (mod)2 Depression Mean (SD) n = 112 Standard 7.9 (3.7) 7.9 (4.0) 6.8 (3.9) 7.0 (4.7) 7.2 (4.0) care n= 107 n=100 n=87 n=75 n=80 9%

18% iCAN-DO 7.9 (3.1) 7.0 (3.4) 6.6 (4.0) 6.6 (3.8) 5.8 (3.8) n=105 n=99 n=78 n=65 n=61 27%*

Anxiety Mean (SD) Standard 9.9 (3.1) 8.8 (4.1) 8.7 (3.9) 9.2 (4.1) 9.4 (3.8) care n=107 n=100 n=87 n=75 n=80 5%

9% iCAN-DO 9.2 (3.0) 7.8 (3.6) 8.0 (3.8) 7.8 (3.1) 7.9 (3.8) n=105 n=99 n=78 n=65 n=61 14%

Complete Depression Mean (SD) cases3 Standard 8.0 (3.9) 7.3 (4.0) 6.8 (3.9) 6.8 (4.5) 7.2 (4.2) 10% n = 112 care n = 62 16% iCAN-DO 8.1 (3.3) 7.3 (3.6) 6.8 (3.9) 6.5 (3.9) 6.0 (3.9) 26%* n = 50

Anxiety Mean (SD) Standard 9.4 (3.2) 8.6 (3.8) 8.3 (3.7) 9.1 (4.1) 9.2 (3.8) 2% care n = 62 10% iCAN-DO 9.2 (2.6) 7.7 (3.6) 7.7 (3.6) 7.6 (2.9) 8.1 (3.8) 12% n = 50

NOTE: High scores in the HADS-D and the HADS-A represent higher levels of symptoms. 1. ITT = In- tention-to-treat, including participants with complete baseline data. Data imputed with last rank carried forward method. 2. ITT (modified) = ITT + at least one more measurement post-baseline. Data were imputed using last rank carried forward method. 3. Complete cases = individuals who completed all assessments and had no missing values. * Above the level of the minimal clinically important difference (according to Puhan et al. 2008). Abbreviations: HADS, Hospital anxiety and depression scale; ITT, Inten- tion-to-treat; SD, Standard deviation.

54 The randomisation tests concerning the HADS-D classifications (non-case, doubtful case, clinical case) showed that a larger proportion in iCAN-DO than in SC changed from doubtful/clinical cases to non-cases (ITT; p = 0.0088). No statistically significant differences were seen regarding the HADS-A. The responder analysis of individuals with complete data on the HADS at baseline and ten months (iCAN-DO n = 61, SC n = 80) revealed clinically meaningful differences in the iCAN-DO group, but not in the SC group, regarding both the HADS-A and the HADS-D (Figure 8). A decrease in depressive symptoms (corresponding to -2 points in the HADS-D) was seen in 33 of 61 participants (54%) in the iCAN-DO group and in 28 of 80 (35%) in the SC group. Mean- while, a deterioration in depressive symptoms (corresponding to +2 points in the HADS-D) was seen in 4 of 61 participants (8%) in the iCAN-DO group and in 21 of 80 (26%) in the SC group. A decrease in symptoms of anxiety (corresponding to -2 points in the HADS-A) was seen in 32 of 61 participants (52%) in the iCAN-DO group and in 30 of 80 (38%) in the SC group. A deterioration in symptoms of anxiety (corresponding to +2 points in the HADS- A) was seen in 10 of 61 participants in the iCAN-DO group (16%) and in 25 of 80 (31%) in the SC group. No significant results within or between groups, regarding anxiety and depression assessed with the MADRS-S, the STAI-S, or the EROS, were seen (Table 7). For the MADRS-S, both the iCAN-DO group and the SC group had mean scores within the levels of mild depression at baseline and ten months, and for the STAI-S, mean scores for both groups were just above the cut-off level at both baseline and ten months. Both groups had unchanged EROS scores at baseline and ten months.

Figure 8. Illustration of the responder analysis for individuals with complete HADS data at baseline and 10 months (iCAN-DO n = 61, Standard care n = 80). Improve- ment was defined as an improvement above the minimal clinical important difference as described by Puhan et al. 2008 (-2 points on the HADS-D and/or the HADS- A). Deterioration was defined as a deterioration of the same magnitude (+2 points the HADS-D and/or the HADS-A). Percent shown on the y-axis.

55 Table 7. Descriptive statistics for MADRS-S, STAI-S, EROS, and PCL-C, at baseline and 10 months, for individuals randomised to iCAN-DO or standard care.

ITT population n = 245 Baseline 10 months Mean (SD) MADRS-S Standard care 16 (7) 15 (8) iCAN-DO 15 (7) 14 (8)

STAI-S Standard care 43 (6) 44 (6) iCAN-DO 42 (6) 43 (6)

EROS Standard care 28 (5) 28 (5) iCAN-DO 29 (6) 29 (6)

PCL-C Standard care 37 (12) 38 (13) iCAN-DO 35 (11) 34 (13)

NOTE: A high score represents higher levels of symptoms/problems, except in the EROS, where a high score indicates higher levels of environmental reward. Abbreviations: MADRS-S, Montgomery-Åsberg Depression Rating Scale; STAI-S, Spielberger State-Trait Anxiety Inventory State subscale; EROS, Envi- ronmental Reward Observation Scale; PCL-C, Post-traumatic Stress Disorder Checklist-Civilian version; ITT, Intention-to-treat; SD, Standard deviation.

Health-related quality of life and post-traumatic stress The results concerning HRQoL, including the specific measures for insomnia and fatigue (Table 8), revealed no statistically significant differences between groups. However, clinically relevant changes were found within and between groups. The mean changes from baseline to ten months suggested improvements of likely clinical relevance regarding role functioning, social functioning, and fatigue within the iCAN-DO group, and of subtle clinical relevance within the SC group. Between-group differences of subtle clinical relevance were also seen, e.g., for social functioning and on several symptom scales (Ta- ble 9). As for post-traumatic stress, measured with the PCL-C, both the iCAN- DO group and the SC group scored below the cut-off level for symptoms at both baseline and ten months.

Table 8. Descriptive statistics for the FACIT-F and the ISI for individuals randomised to iCAN-DO or standard care.

ITT population n = 245 Baseline 1 month 4 months 7 months 10 months Mean (SD)

Standard care 31 (11) 31 (11) 34 (11) 35 (11) 34 (12) FACIT-F iCAN-DO 31 (10) 30 (9) 33 (11) 35 (10) 35 (11)

Standard care 12 (6) 13 (6) 12 (7) 12 (7) 12 (7) ISI iCAN-DO 11 (6) 11 (6) 11 (7) 11 (6) 11 (7) NOTE: A high score in the FACIT-F represents a low level of symptoms/problems. A high score in the ISI represents a high level of symptoms/problems. Abbreviations: FACIT-F, Functional Assessment of Chronic Illness Therapy Fatigue Subscale; ISI, Insomnia Severity Index; ITT, intention-to-treat; SD, standard deviation.

56 Table 9. Descriptive statistics and change over time for individuals randomised to iCAN-DO or standard care.

ITT population n = 245 Baseline 10 months Mean clinical Clinical differences Mean (SD) changesa between groupsb

Global health status Standard care 47 (22) 54 (23) +7 S Trivial

iCAN-DO 48 (21) 56 (22) +8 S

Functional scales

Physical functioning Standard care 80 (18) 81 (19) ND Trivial iCAN-DO 77(21) 82 (20) +5 S

Role functioning Standard care 59 (34) 70 (32) +11 S Trivial iCAN-DO 52 (32) 75 (32) +23 M

Emotional functioning Standard care 56 (21) 59 (25) ND N.A. iCAN-DO 60 (18) 64 (23) ND

Cognitive functioning Standard care 64 (28) 65 (25) ND iCAN-DO 64 (26) 67 (24) +3 S Trivial

Social functioning Standard care 60 (29) 65 (28) +5 S iCAN-DO 58 (29) 70 (26) +12 M Small

Symptom scales/items

Fatigue Standard care 48 (25) 43 (26) -5 S Trivial iCAN-DO 52 (26) 40 (25) -12 M

Nausea and vomiting Standard care 11 (14) 7 (12) ND Trivial iCAN-DO 14 (19) 9 (12) -5 S

Pain Standard care 30 (28) 26 (26) ND Small* iCAN-DO 34 (29) 33 (28) ND

Dyspnoea Standard care 27 (28) 33 (30) +6 S Small iCAN-DO 35 (31) 35 (29) ND

Insomnia Standard care 45 (32) 47 (34) ND Small

iCAN-DO 42 (31) 40 (32) ND

Appetite loss Standard care 19 (27) 17 (24) ND Trivial iCAN-DO 25 (30) 14 (22) -11 S

Constipation Standard care 17 (26) 16 (27) ND Trivial

iCAN-DO 14 (25) 14 (25) ND

Diarrhoea Standard care 18 (25) 12 (22) -6 S Small* iCAN-DO 16 (27) 16 (25) ND

Financial difficulties Standard care 21 (29) 17 (25) -4 S Small iCAN-DO 26 (34) 21 (33) -5 S

NOTE: A high score for the global health status represents a high QoL. A high score for a functional scale represents a high level of functioning. A high score for a symptom scale/item represents a high level of symptoms/problems. a. Mean clinical changes within group from baseline to 10 months. For QLQ-C30 scales, as defined by Cocks et al. 2012: ND = no/unlikely difference, S = small change of subtle clinical relevance, M = medium change of likely clinical relevance b. Clinical differences between groups at 10 months. For QLQ-C30 scales as defined by Cocks et al. 2011: Trivial = no difference between groups, Small = a difference between groups of subtle clinical relevance. * Standard care advantage. N.A. = no available guideline. Abbreviations: EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; SD, standard deviation.

57 User experiences of relevance and benefit (Paper III) The content analysis resulted in four main categories: “Gaining knowledge and support but wanting more personalisation”, “A feeling of safety that was needed earlier”, “Own situation, preferences, and timing determine the use of peer support” and “A complement to standard care”, abstracted from 17 subcategories (Table 10). The results are presented in a structure based on the categories, the second and third category are presented together.

Table 10. Categories and sub-categories from the qualitative content analysis. Categories

Gaining knowledge A feeling of Own situation, A complement to and support but safety that was preferences, and standard care wanting more per- needed earlier timing determine sonalisation the use of peer support

Sub-categories

The importance of Reliable and Not interested in The information information and trustworthy peer support at all, given at the hospi- support besides lurking tal was insufficient

Confirmation, Wanted to have Higher presence iCAN-DO as a recognition, and be- access to Step 1 of healthcare pro- source of infor- ing taken seriously earlier fessionals in the mation for others forum

A wish for more Seeking infor- Facebook pro- Standard care did specifically tailored mation online vided a better en- not offer any sup- contents began at an early vironment for port for emotional stage online peer sup- problems port Positive, supportive Information contacts with the needs varied Fit into everyday psychologist over time life

Turning down the offer of iCBT

Limitations of iCBT

58 Gaining knowledge and support but wanting more personalisation In this category, the importance of information and support was highlighted by informants, who used iCAN-DO to inform themselves and increase their understanding of what they were going through. The information gathered and increased knowledge resulting from this were described as a way to cope with the situation. Some of the informants who had used both steps 1 (IAS) and 2 (iCBT) in iCAN-DO said that the steps complemented each other: Step 1 was seen as a source of reliable information that served as a foundation for further processing with the help of Step 2. Step 1 was perceived as calming and con- firmatory and it was valuable to users when they could recognise descriptions of symptoms. The contents were not described as intimidating by any of the informants; on the contrary, the informants expressed relief at seeing that the contents matched their reality (e.g., recognising bothersome symptoms). The recognition factor was described regarding the library content, the “Ask an expert” section, and the peer support section.

“I asked some questions myself, but I have also read the questions that others have asked and the answers they received. I wanted to check if there was anyone who was in the same position as myself. Is this I’m feeling normal? That recognition factor has been important to me.”

Female, Breast cancer

The section where informants could pose questions was appreciated by those who had used it and was seen as educating and encouraging. However, the function would have been even more valuable if the nurse who answered the questions had had access to personal medical records when providing the answers. It was suggested that “Ask an expert” would be suitable to implement in regular healthcare. Other suggestions from the informants included contents more tailored to individual needs. Some of the factors mentioned as important were diagnosis, age (feeling much younger than other patients could create a sense of loneliness), gender, treatment, and symptoms.

“If I don’t feel that the content concerns me, then I don’t want it…Like these things concerning sexual problems…it must concern me individually, for example be based on the type of treatment that I have received.”

Male, Colorectal cancer

The informants who had worked actively with the iCBT programme experienced positive, supportive contacts with the psychologist and stated that the

59 material was helpful in managing symptoms. The programme was described as demanding, but worth the effort. In contrast, the demands could feel overwhelming, e.g., when having to decide what modules to work with.

“I felt that I needed every module. This wasn’t for me… perhaps if my life situation had been different. This would have been good for me, but it didn’t work. It became so stressful in a situation where I couldn’t handle any stress.”

Female, Breast cancer

Among those who used Step 2, the experiences of working with the psychologist through writing online varied. Some, who were satisfied with the level of interaction, saw this as the preferred way of communicating. Others described the psychologist as absent and faceless in this form of therapy. Some suggested that video meetings could work as a facilitator. There were also suggestions that the programme should had been integrated into standard care and visible in personal medical records.

Some informants explained why they did not continue with the second step. The feeling of sufficient support from relatives and friends, a fear of making it worse by focusing on negative feelings, and lack of confidence in iCBT treatments were mentioned.

A feeling of safety that was needed earlier, and the use of peer- support Step 1 was seen as reliable and trustworthy, compared with other online information. It could be difficult to verify online information from other sources and it was considered a relief to be able to trust the content. Another aspect of safety was availability. Even if information and support was not needed at a certain time, it felt reassuring to know that it was easily available.

“Well, it gives you a feeling of safety, that this is what is considered accurate in the world of research-this is the advice we want to give. Like, for example this regarding physical activity, I felt that ok I will continue with my exercise because they say it’s good. It is like a support. Some things I might already know about, but here I can get confirmation that it is accurate.”

Female, Breast cancer

The timing of the intervention was highlighted by most, who stated that they would have wanted access to Step 1 earlier, at the time of diagnosis. At the time of diagnosis, information needs were very high and the search for online information could start as soon as they got home from the hospital, or even

60 before that, during the investigation phase. The online information found outside iCAN-DO was described as sometimes helpful and correct, but it could also be inaccurate, scary and eliminate hope.

“The same day I got the news of my poor prognosis from the oncologist, the first thing I did when I came home from the hospital was to go online to search, and I ended up on personal blogs…I call them death blogs. I don’t know…there are descriptions of terrible side effects. In some cases, the blog ended with next-of-kin writing an obituary over the person who was now deceased. All that is left online for the rest of us, I wish it wasn’t like that because when I read it, I felt like…everyone dies.”

Female, Breast cancer

The informants interested in peer-support had already found their platforms when getting access to Step 1, some had accessed Facebook groups and the descriptions of their helpfulness varied. Most informants were not interested in active participation in the peer-support section, but lurking (non-active participation) was common. It was suggested that a higher presence of healthcare staff in forums would ease participation.

A complement to standard care Step 1 was described as a complement to standard care. It could be used when the information given at the hospital felt incomplete, e.g., was difficult to understand or was given quickly due to limited time. Informants also described how being in a state of shock limited their ability to remember what was said at the hospital; Step 1 could then be used as repetition. The fact that there was consistency in the information from the hospital and Step 1 was described as a relief. Factual information and self-care advice were seen as highly appropriate to deliver in web-based form, but some participants stressed that emotional support should be provided face-to-face. Further, both Step 1 and Step 2 were described as means to inform others. Family members’ information needs were high as well, and it could be perceived as burdening to inform them. To be able to read the same information together was considered valuable.

“Because when you get this kind of news you become totally blocked. It went somewhat ok for me to listen to what the physician said, but it was so brief. The physician that informed me was quite brief, at least I felt that. It was like he just reported a finding, not so much information so…so it became important to me to find information from other sources.”

Male, Prostate cancer

61 Lastly, it was described as difficult to fit standard care into everyday life. It was described as hard to get in touch and to get adequate answers to questions. The high accessibility in iCAN-DO was seen as positive by informants, who suggested several possible implementations. iCAN-DO was intended as a complement to standard healthcare, but in addition to the problems of getting access to standard care, several examples of unmet needs and dehumanising care were described by the informants. This could potentially have formed a fifth category “unmet needs and dehumanising standard care”, though this would not have fallen within the scope of Study 2. One example, describing an emergency room visit due to the cancer, is provided below.

“It was such a lacking organisation. There was a lack of planning, but most of all there was a lack of humanity! I saw one woman, another patient, she was in severe pain. Then the physician came, and right there among all the other people she undresses this woman, naked, to examine her. Then she came over to me, because it was my turn and I said right away: ‘Stop, what are you doing?’ ‘I am going to examine you,’ she said. Then I said: ‘But we are right out in the middle of the corridor.’ ‘Oh, I’m sorry,’ she said, and rolled me into an examination room, but this was only because I protested. This was last year; one would think that healthcare would had come further. This young physician, it was like she woke up from some kind of blunted state when I protested, and then she finally saw me, but only because I protested. I am used to saying what I feel, but the other woman, the one who was undressed in the corridor… Who will speak for her?”

Female, Breast cancer

62 User experience of delivery, design, and structure (Paper IV) The content analysis resulted in the theme “The cancer disease and its treatment place high demands on accessibility and user experience”, based on three subthemes and ten categories (Table 11). The results are presented in a structure based on the subthemes.

Table 11. Sub-themes and categories from the qualitative content analysis. Sub-themes

User experience in the Technical struggles re- Appealing and usable, context of cancer quire patience and trou- but rather simple bleshooting

Categories

My health status and en- Complicated multistep Intuitive but instable in- vironment affect how I login procedure terface use the portal

Multiple delivery modes Annoying technical Appealing yet simple enable assimilation of problems reduce motiva- graphical design the content tion to use the portal

A web portal provides Troubleshooting and the Technology that creates high availability and ac- importance of support motivation and innova- cessibility tion

Questionnaires make you reflect on your health

User experience in the context of cancer All informants mentioned their personal situation and the consequences of the disease, both somatic and cognitive, as factors that influenced their abilities to make use of iCAN-DO. Symptoms like pain and fatigue affected use, e.g., needing to lie down or use a lightweight device instead of a computer due to pain. Further, some informants suggested shortening or dividing the video

63 clips, since it was difficult to pay attention for too long. Non-use due to symptoms was also described.

“No, you are tired and sometimes it just doesn’t work, it’s like your head doesn’t work, you don’t have the energy. It can be due to the cancer, but also other problems, the cancer is just one of many problems…”

Male, Prostate cancer

“I was tired as well, so I didn’t really have the strength to look at it. Sometimes I fell asleep…”

Male, Colorectal cancer

Being able to use the portal on mobile devices as well as on a computer was seen as important. Further, informants described a variety of preferences regarding information formats. Some assimilated the content best by reading, others through audio or film, or a combination thereof. Some printed all the material and read it on paper, and a few described writing their own notes on them. One of the informants showed a thick binder in which major parts of the material from the portal had been collected in printed form. This was where the informant read, without logging in to the portal. Still, the high availability of iCAN-DO was appreciated by the informants, who described the ongoing oncological treatment alongside the regular exertions of everyday life as highly stressful. Further, answering questionnaires in the portal was described as being both helpful and trying. When the informants noticed an improvement in health, this felt good, but when the opposite occurred, or when they were reminded that certain symptoms might occur, this could feel discourag- ing. One informant stated that questions regarding sexuality were particularly trying.

“There was a recurring question about whether I had sex or not, and that really bothered me, because you feel so damn bad as it is! It was not even relevant to me, and it felt bad, like it’s none of your business. The worse I felt, the more frequent that question was.”

Female, Breast cancer

Technical struggles require patience and troubleshooting The two-step verification used for login was experienced as difficult and some users had problems receiving the codes. The process was described as some-

64 thing that decreased the number of spontaneous visits. Several informants suggested an implementation of Mobile BankID instead. Other technical difficulties described by informants were the need for extra add-ons to make the page work, a perceived instability causing involuntarily logouts, or having problems communicating with the psychologist. Examples of the latter could be difficulties understanding where in the system the answer would be found, or that the answer still looked unread even after having been read. The technical difficulties led to postponing visits and lack of motivation in using the system.

“Because otherwise if you don’t get help you feel like you don’t give a damn…When there are technical difficulties on the site you postpone your visits, I will log in later…but then you never do it.”

Male, Prostate cancer

Still, informants had several problem-solving strategies, e.g., installing new software or contacting a relative or friend that could help them. One commonly described strategy was contacting the U-CARE helpdesk for technical support. Getting technical support was seen as crucial to not giving up.

Appealing and usable, but rather simple iCAN-DO was easier to use than many other systems and what could be gained from the contents was seen as more interesting than having an advanced interface. Some informants found the interface intuitive and easy to understand and navigate, while others experienced this simplicity as a draw- back. The graphical design was described as simple and serving its purpose, but some found it dull and intricate. These experiences varied between informants who were experienced everyday computer users with a high interest in social media and informants who were inexperienced with no or low use of social media.

“I am not into computers at all, but I think I could manage well with this anyway and that must be a good review.”

Male, Prostate cancer

“I think it’s good and simple, it doesn’t have to be fancy if it is genuine, and the content, like the library where you can read about stress and anxiety and things like that, I think it is very good. It helped me. And I am not active in…like social media and so on, I’m not a Facebooker at all… if you see what I mean. I am only interested in the facts, that’s all I want.”

Female, Breast cancer

65 “It was annoying to use really…and well, I lost interest. I’d much rather hang out on Facebook instead.”

Female, Breast cancer

“I am into Instagram, Facebook, and all those places you know, so I felt this was…a bit boring. I am sorry to say but that’s just me. It doesn’t give me the kind of input and variation I need.”

Female, Breast cancer

Additional experiences regarding the usability concerned the reminders. When a participant in AdultCan had questionnaires to answer in the portal, repeated reminders were sent via SMS, and this approach was not reported as disturbing by anyone. Informants experienced the reminders as helpful, not only in answering questionnaires but also as a trigger to make use of the supportive material, when they were logged in anyway. Finally, informants provided several suggestions of improvements, e.g., the layout of questionnaires, showing results from the self-reports as graphics, and adding updates concerning the latest news in oncology research.

66 Discussion

Main findings The overall aim of this thesis was to explore and evaluate design, delivery, and effectiveness of internet-based stepped care for individuals with cancer and self-reported symptoms of anxiety and/or depression. Describing the development in the initial steps of this research served as an opportunity to gain a deeper understanding and identify possible explanations when interpreting the effects and experiences of using the intervention (iCAN-DO). This increased the possibilities of recommending components that might be of value in similar interventions. Results from the RCT suggested that iCAN-DO improved symptoms of depression, while anxiety and secondary outcomes were largely unaffected. Further, participants used the low-intensity interactive support (Step 1) to a higher extent than they used the second step, iCBT. The qualitative exploration conducted alongside the RCT suggested that the development including co-creation and theoretical input added relevance to Step 1. However, the experiences of lacking personalisation and issues with design and structure indicate that there is room for improvements. The results may also be used to explain other observations, such as attrition problems, and to suggest the focus of future research.

The development of interactive support (Paper I) We set out to develop Step 1 using the evidence-based triad (161) and theory- based input. This resulted in a low-intensity support guided by the concepts of psychoeducation (111), the S-CDNT (118) and the SCT (162), containing factual information and suggested self-care measures provided in different forms. Since hopelessness is an important factor in the development of depression, strengthening the user’s knowledge, abilities, and strategies seemed important and the goal was to empower users by providing a relevant self-care agency. Already in the early stages of the development process, it became clear that several mistakes could have been made if the chosen approach of co-creation had not been used. When identifying potential self-care deficits/demands, the input from both the clinician group and the lived experience group was highly useful. Further, the need for a clear connection to cancer in all materials was highlighted at an early stage by the lived experience group, e.g., a general

67 lecture on anxiety was not seen as useful as a lecture on anxiety in relation to the cancer. Thus, it seems important to recognise that the symptoms/problems experienced are not isolated events, but closely related to the cancer or the situation that the cancer creates.

Another important issue was that of the delivery method. Researchers and clinicians are familiar with written text as a mediator of knowledge. This may not be the case for all of the intended users, who may in addition be suffering from symptoms such as fatigue, known to affect attention, concentration, learning, and memory (175). These facts were stressed by the lived experience group and resulted in all the lectures being provided as both written text and audio-visual presentations.

Working with co-creation may also lead to a number of challenges. Some challenges described within past collaborations of this kind are the imbalance of power and conflicts of interest (176) that may be present. From our perspective, cooperation with both groups was positive and highly valuable. Still, some aspects of the cooperation can be discussed. In the beginning of the process, we provided the lived experience group with basic knowledge on the planned research process. However, the group could have been provided with additional knowledge on research principles, since it would have balanced the power structure and increased the group’s possibilities to contribute further. In addition, the lived experience representatives were recruited from patient associations. This was positive in several ways as they were both strong, engaged, and knowledgeable, with a high interest in contributing to help others with “their” diagnosis. Still, far from all people with cancer will engage in patient associations and the views of those who do may not be generalised to all intended users. The adherence in the iCAN-DO group in AdultCan suggests that the intervention appealed most to those with a higher level of education. Thus, it would probably have been valuable to involve representatives with different backgrounds, ages, and educational levels, to get a varied input. It also seems important to recognise that power structures may be present within all groups, and one suggestion to make it easier for all individuals to be equally heard is the use of anonymous voting processes as part of decision- making (176).

As stated earlier, the triad included the research team, clinicians, and individuals with lived experience. In retrospect, one group that was missing in the constellation was the system developers. During the development process, we invited representatives from this group, but cooperation mainly consisted of the research group reporting findings and results from discussions. In order to maximise the chances of achieving high usability, it would likely have been valuable if system developers had been present, instead of them getting second-hand reports from research group members without knowledge on system

68 development, including the specific language, research methods, and relevant outcomes. Collaborations to increase usability of a system, called co-design, have been shown to increase the acceptability of systems and may include focus groups and interviewing as well as recording of sessions when the system is used (177, 178). Adding codesign to the development would most likely been valuable.

The effects of iCAN-DO (Paper II) Effects on anxiety and depression The screening procedure identified 245 individuals (27%) with symptoms of anxiety and depression and this is in accordance with a previous study in a similar setting (38). The ITT analysis revealed statistically significantly lower levels of depression symptoms in iCAN-DO than in SC (LMM), and the analysis of the HADS classifications (randomisation test) showed a significant decrease in the proportion of doubtful/clinical cases of depression in iCAN-DO, but not in SC. These positive results were partly supported by the sensitivity analysis. In those results, clinically significant decreases of depression were seen within iCAN-DO, but not within SC, in all populations. However, no statistically significant differences were found in the sensitivity analysis. In addition to the decrease in symptoms of depression, the responder analysis showed that a larger proportion in SC than in iCAN-DO deteriorated in symptoms of both depression and anxiety.

Most participants used the low-intensity interactive support (Step 1) and the positive effects on symptoms of depression reveal the value and importance of providing supportive nursing care in various forms within oncology. The previously mentioned Cochrane review on psychosocial interventions (99) concluded that nurse-delivered interventions, including psychoeducation, had a positive impact on mood in newly diagnosed individuals with cancer. Since none of the interventions included in that review were internet-based, the findings from Paper II contribute to the picture. Another aspect of the results on symptoms of depression that may be regarded as positive is that they were seen in the long-term follow-up (10 months). The previously mentioned systematic review on self-management interventions for individuals with cancer (114) found some web-based interventions (179-181) with initial effects on symptoms of depression. However, these effects did not remain at followup, or were not investigated.

The positive results in symptoms of depression were related to the HADS scale. No similar changes were seen on the MADRS-S, a fact that needs to be

69 addressed. The main reason for including the MADRS-S was to identify participants in need of further psychiatric care. Ten individuals were identified in this way, but none was assessed as needing referral to psychiatric care. A possible explanation why no effects were seen on the MADRS-S are its psychometric properties, based on the diagnostic criteria (DSM) for depression; it is designed to detect changes in primary depression during treatment with anti- depressants. Meanwhile, the HADS was designed for individuals with an ongoing illness. One study focusing on older females with advanced cancer comparing psychiatric clinical interviews with PROMs found a high agreement between the MADRS-S and the DSM, but no agreement between the HADS and the DSM in that population (182). Thus, it can be suggested that the scales measure different aspects of depression. There is a lack of convincing studies regarding the feasibility of the MADRS-S in individuals with cancer, and some of its questions clearly cover common symptoms during oncological treatment, e.g., appetite and concentration. Therefore, it is difficult to validate the results and further studies are needed to determine the feasibility of the MADRS-S in the cancer context. The EROS has been used in the cancer context in trials concerning CBT (183), but no studies on self-care interventions using the EROS as an outcome measure could be found. The fact that iCAN- DO involved very few individuals receiving the second step of iCBT may have contributed to the null effects for this outcome measure.

Regarding anxiety, results were similar across analyses, and no statistically or clinically significant results were seen. This suggests that iCAN-DO affected symptoms of depression rather than anxiety. Fewer participants used the content targeting anxiety, though more participants were included on the basis of > 7 points on the HADS-A than on the HADS-D. The constructs of anxiety in this population are most likely a natural response to the real threat implied by the cancer and may therefore be looked upon by the participants as something difficult to counteract. Symptoms may also be diffuse and not easily recognised as anxiety e.g. sleeping problems, irritability or restlessness. Some interventions use tracking of symptoms to recommend appropriate materials and self-care measures to a user (184); this may add value for some individuals, but not all. It is important to recognise that symptom tracking may cause increased distress in individuals, e.g., those who initially cope with their cancer by avoiding information (185). In addition, it would have been valuable to have input from individuals with lived experience of anxiety and depression during cancer in the development phase.

The results may also indicate that iCAN-DO was not tailored well enough to target anxiety. At the ten-month follow-up, anxiety may have been fuelled by fear of cancer recurrence (FCR) (40). Those most at risk for developing FCR have been suggested to be individuals with breast cancer, younger age, a history of generalised anxiety disorder, or previously stressful or traumatic life

70 events. A systematic review found the frequency of FCR to be 39–97% across cancer types and assessments (186); among women with breast cancer, as many as 70% reported FCR (187). In this last study, time since diagnosis was not shown to be statistically significant; another study suggested that prevalence rates were highest between 1 and 2 years post-surgery, and then decreased, although they were not nil even after 12 years (188). Thus, FCR is a common, long-term problem and in future interventions it would most likely be of value to target the underlying mechanisms specifically. For instance, a previous RCT involving psychoeducation and CBT showed the potential to target FCR specifically (189).

Effects on HRQoL and post-traumatic stress The results (ITT) for HRQoL, including specific measures of insomnia and fatigue, showed no statistically significant differences. However, within- group differences of likely clinical relevance were found in the iCAN-DO group, but not in SC. Between groups, clinically relevant effects of subtle relevance, with greater improvement in iCAN-DO, were seen regarding several symptom scales, as well as in social functioning. Returning to the recent review of self-management interventions for individuals with cancer (114), these show mixed results for HRQoL. Of 26 studies reporting on HRQoL, 15 reported statistically significant improvements, 10 showed no changes, and one reported worsening. Most (n = 11) of the studies with positive results had a primary endpoint ≤ 6 months. Only one of the interventions was internet- based, targeting men with prostate cancer and their spouses, with its primary endpoint at 8 weeks (190). Thus, it was difficult to make any comparisons with the results from iCAN-DO; more studies on internet-based self-care interventions with long-term follow-ups are needed. The results for post-traumatic stress showed that both the iCAN-DO group and the SC group scored below the cut-off for symptoms at study start and end. Therefore, little can be said about the appropriateness of iCAN-DO in post-traumatic stress. In addition, the PCL-C and the PCL-5 are no longer seen as the most suitable instruments in the cancer population which may also explain the findings.

Did the intervention happen? Step 1 was used (i.e., material opened) by most of the participants, while Step 2 was used by very few. However, it is important to note that definitions of adherence in internet-based interventions may include, e.g., number of logins, time spent, number of clicks, number of openings, or recordings of activities. In iCAN-DO, the measure was based on number of times a material or section was opened. This is naturally imbued with some uncertainty, as there is no way of knowing what happened next. Further, if a participant chooses to print a material and read it on paper, the count is lost. In Step 2, the psychologist

71 could monitor the participants’ progress within a module with somewhat greater certainty (based on fulfilled assignments). As stated, definitions of usage and adherence in internet-based interventions are not uniform. A systematic review (155) also suggested that the individual user’s goals should be included in the definition, and since these may differ, all users may not have to access the same contents or features to achieve their goals, i.e., more use is not always better. In iCAN-DO, usage patterns can be studied in greater detail in future studies. Since previous results indicate that far from all clinically recruited participants with symptoms of anxiety and depression are in need of psychological treatment (38), stepped care as a model for the intervention seems relevant. However, anxiety and depression were still present in individuals at study end and the low adherence to iCBT in combination with the results from Papers III–IV suggest that further efforts are needed to increase efficacy and to make iCBT feasible in this context.

Perceived relevance, benefits, and aspects of usability (Papers III–IV) User experience of relevance and benefit The overall impression was that informants had substantial unmet supportive needs that were partly met by the contents in iCAN-DO. iCAN-DO was seen as a reliable supplement to, or sometimes replacement of, standard care. The appropriateness of the development and content in Step 1 was supported by the informants’ descriptions of how they used it and what purpose the information gathering had for them. Information was described as central to understanding and managing the new situation, in good agreement with the theories used as guidance during development. The need for information was described by most as being largest at the time of diagnosis and several had searched online early, some even before being diagnosed. Informants described difficulties in verifying unregulated online content and there were also descriptions of fearsome and unsupportive environments online. Thus, iCAN-DO was a valuable alternative, which was seen as trustworthy. Still, some of the informants described an inability to initially absorb any more information, and the knowledge that iCAN-DO would be available for a long period of time was then seen as advantageous. Information-seeking behaviours are likely connected to coping strategies within the individual (185). For some, early information-seeking is a way to make sense of the situation and take back control, while some individuals will initially want to block information out (191). This suggests that internet-based support should be available not only in the beginning, but over a long period of time. In addition, most of the informants discussed the timing of the intervention and would have wanted access earlier, preferably when diagnosed. During the development phase, a decision was

72 made not to include participants on the day that they received the news of the diagnosis. However, the time of diagnosis is most likely a suitable occasion for implementing internet-based support.

Information-seeking behaviours in individuals with cancer are described as including a need for narratives from others (191), often used as a way to understand the own situation. Sharing such descriptions can be positive, but may also lead to misconceptions and fear. Being able to read about the experiences of others in iCAN-DO was appreciated by informants, who perceived this as confirmation that they were not alone. Informants described reading others’ questions in the FAQ and the peer support section. Most were satisfied with lurking and did not want to interact with others. This is very common, often described as the 1% rule (192). The rule suggests that 90% of users in internet forums are lurkers/observers, 9% will contribute seldom, while 1%, the “superusers”, will account for the majority of the content. Thus, superusers are important, but when they are the only contributors this may distort the emerging picture of a situation or problem. It has been suggested that superusers may be those reporting higher levels of symptoms, while lurkers may be more active in working with the contents of an intervention (192). It seems important to create an environment where lurkers feel safe enough, or motivated, to contribute. Security and reliability have been suggested as important to lurkers (193) (along with personal characteristics, e.g., shyness and perceived self-efficacy) and the informants mentioned that a higher visibility of healthcare staff within iCAN-DO forums would have made them more likely to contribute. Thus, when forums are used in interventions, moderating should include interactions not only in case of problematic behaviour, but also a regular visibility and the creation of a supportive, reliable, and safe environment.

Informants suggested that a higher degree of personalisation of the contents in iCAN-DO would have been valuable. Still, the content in Step 1 was seen by informants as suitable to integrate into standard care. Such integration, e.g., connection to personal medical records, might lead to increased possibilities to personalise the contents (194). In the present study, connection to personal medical records was not deemed possible, and the lack of interoperability between systems within healthcare is a major limitation when developing internet-based interventions. Hopefully, the forthcoming “My Care Plan Cancer” (137) is an integrated step in the right direction. For example, the contact nurse may communicate online with patients, as well as being able to suggest relevant modules of information for the individual in question. All informants in the present study experienced that information and self-care advice were highly appropriate to deliver online, while some suggested that emotional support should be delivered face-to-face. iCBT was experienced as helpful but demanding, as the informants were in a situation that was already strained. The limited time for therapy (10 weeks) was a source of stress, as was deciding

73 what modules to work with. The latter was due to a feeling of “needing them all”, as well as experiences of symptom interaction, i.e., symptoms were hard to separate from each other and match to the modules. Current symptoms and time constraints have previously been described as hindering iCBT for individuals with cancer (195). Since cancer treatment might cause a high symptom burden, the acceptability of similar programmes needs to be carefully explored in future interventions, i.e., by including individuals with lived experience of cancer and using feasibility studies. Informants further suggested that it would be helpful to have a more personal contact with the psychologist at the start of the programme, e.g., by a video meeting. A recent randomised trial on short self-guided iCBT targeting individuals with cancer (196) included one arm with weekly guidance; participants could choose if the guidance would take place through a telephone call or e-mail contact. Results illustrated that guidance was advantageous, with more completed modules and higher treatment satisfaction in the guided group compared with in the self-guided group. Thus, guidance served a purpose and being able to choose the preferred way of contact might also be valuable. Using blended iCBT to overcome some of the barriers might be another solution. In the blended therapy form, some sessions are delivered face-to-face and some are delivered via the internet (189). The internet-based therapy form may be an advance for individuals during strenu- ous treatments, but the program needs to be adaptable to the individual regarding e.g. content and duration.

All in all, it was seen as a privilege to have access to iCAN-DO, as it was difficult for informants to get support and answers to their questions within standard care. The trustworthiness was an important factor, since the quality of other online information and support varied. Thus, using the TAM vocab- ulary, the perceived usefulness (157) was evident. The flexibility of iCAN- DO was appreciated and the possibilities to share the information with others, e.g., family members, was highlighted. Next-of-kin are an important group and in the role as informal caregivers they are of high value, not only for the individual with cancer, but also for healthcare providers. Being an informal caregiver is complex and when overburdened, caregivers may also experience effects on their health, e.g., developing symptoms of anxiety and depression (19). Since internet-based interventions have the potential to decrease mood symptoms in next-of-kin (197) it would be beneficial to include them more specifically in future interventions. Naturally, these interventions also need to be co-created with informal caregivers to add relevance (198).

“The fifth category” The informants’ needs to talk about their feelings relating to the cancer during the interviews were expected. However, the examples of unmet needs and dehumanising care were not as expected. The Health and Social Care Inspec-

74 torate’s definition of irregular events encompasses events that cause the individual psychological damage; some of the quotes described events that perhaps should have been reported as irregular. In addition, it is impossible not to wonder how much impact the events had on individual development of symptoms of anxiety and depression. How did they affect the participants’ sense of hope? The blunted state among healthcare staff described by one of the informants is most likely caused by the strained situation in hospitals but remains unacceptable. This is acknowledged in a recent article by Richards and Borglin (199), describing it as a form of neglect of the professional responsibilities within nursing. The problem is not solely that of nurses and concerns all categories of staff working within healthcare, but it is each group’s responsibility not to become blunted in the performance of their profession. It is crucial that nurses and teachers within nursing are aware of and work consciously to avoid “shitty nursing” (199). The article further identified a major lack of experimental research investigating the impact of interventions in nursing care.

User experiences concerning usability The perceived usability of iCAN-DO was clearly affected by the informants’ ongoing cancer and treatment. They experienced both physical and cognitive symptoms in combination with everyday stress. The positive aspects of iCAN- DO included its flexibility and just-in-time access. The use of different formats in the delivery of information and support also stood out as positive, and a variety of preferences was reported. Offering several formats has previously been reported as beneficial, especially for underserved groups (200). During interviews, it was noted that some informants experienced the didactic information as most useful, while some reported narratives as their preference. However, no analysis of interaction between the chosen material and demographics has yet been conducted. As earlier stated, there are uncertainties concerning the definitions of usage in iCAN-DO; thus, in future interventions, recordings of participants’ activities could be recommended (201). Further, iCAN-DO did not incorporate responsive design in the beginning – it was a purely computer-delivered intervention. Informants stressed the need for re- sponsivity, which was later incorporated.

It is known that participants in studies may perceive negative aspects of PROMs. Questions may work as negative reminders or triggers for new trou- blesome thoughts. The results indicate both negative and positive aspects of the PROMs and some informants mentioned that it would be valuable to get their symptoms visualised graphically. Symptom tracking may be used as feedback to participants (self-monitoring), as a mean to suggest materials based on reported symptoms, or as determining when healthcare staff should

75 contact a participant (202). However, the use of self-monitoring is not unprob- lematic, and users have been reported to game the system (to avoid being contacted) or being negatively affected, with increased levels of anxiety and depression (184). Thus, implementation of self-monitoring should be done with some caution and may not be best used as default, but rather as a possible choice for the individual. Further, if implementing such functionality, it seems important to include a way for participants to get advice and feedback concerning the results.

Results revealed technical difficulties, leading to negative emotions, a decrease in spontaneous visits, and postponing behaviour. Thus, ease of use as described within TAM (157) was not evident and technical support was seen as crucial to not giving up. In retrospect, extended usability testing could have been performed before launching the intervention and, as stated earlier, system developers could have been more included and present during development. A factor that may have been important for the survival of the intervention was that the contents in iCAN-DO seem to have been desired by the users, thus increasing endurance in dealing with technical difficulties (157). Further, informants engaged in various problem-solving strategies to increase usability. Mobile BankID has been implemented, as suggested by informants, which drastically reduced the need for support due to login problems.

The user experience was perceived differently by experienced internet users, with high interest in social media, and those less experienced, with no or low use of social media. This might be explained by the development process and composition of the lived experience group, consisting of individuals who all stressed the need for simplicity with regard to design and features. In retrospect, it would most likely have been beneficial to set up a larger, more heter- ogenous group (or several smaller groups). Thus, future interventions would gain from a more meticulous user analysis and including input from those with lived experience from different backgrounds, as well as with differing use of internet and social media.

Methodological discussion Study 1 The appropriateness of the methodology in Paper I and suggestions for future interventions have largely been covered and will not be discussed further. As for the intervention itself, several issues have been discussed, the main strength being the evidence-based development including co-creation, which is often missing in similar interventions (114). Concerning Paper II, there are several methodological considerations left to discuss. As regards strengths,

76 the RCT design, including screening (reducing the risk of floor effects) and concealed randomisation, should be mentioned, along with the experimental design and long-term follow-up. Further, the trial was registered (clinicaltrials.gov NCT-01620681) and reported according to the Consort statement (203). Regarding socio-demographics and medical characteristics, the iCAN- DO group and SC group were alike at baseline, but it can be discussed to what extent a selection bias occurred. Internet-based interventions often report recruitment taking place through advertising and self-referral (144, 204), naturally creating a selection bias. The clinical recruitment in AdultCan can therefore be seen as a better option. Still, results showed that those declining participation were older and more often diagnosed with colorectal cancer, suggesting that the results may be restricted to middle-aged individuals with a favourable prognosis, i.e., females with breast cancer. Further, a common reason to decline participation was the lack of computer or internet access. This might be by choice, but some presumptive participants stated that they could not afford a computer. It is well-known that lower socio-economic status and education, comorbidities, and being part of a minority entail more difficulties in healthcare contacts (205). In addition, psychological problems may be more stigmatising among blue-collar workers than in other parts of society. Since individuals with cancer and shorter education are also predisposed to depression (206), these groups are especially important to reach. In CHESS (147), where the focus was to include underserved women, computers were installed in the participants’ homes during the study period and participants were also provided with training on how to use them. While such an approach should be carefully considered regarding appropriateness, it is highly important to con- sider how to better include the underserved in future studies (207). A responsive design to allow for several types of devices may be beneficial. The fact that individuals with lower levels of education did not remain in the iCAN- DO group to the same extent as those with longer education also strengthens the importance of including input from individuals with different backgrounds in the lived experience groups when developing interventions. In Paper II, the differences seen in level of education between groups at ten months led to all analysis being adjusted for this.

PROMs were used to explore the individual perceptions of participants. An obvious advantage of the internet-based setting is that all answers were filled in electronically, thus being concealed and never overlooked by the research team. Further, the primary outcome measure, HADS, is validated for the population and relevant for the purpose of screening (90). However, a thorough clinical assessment has been shown to identify more individuals with anxiety and depression than the HADS and it should be taken in consideration that males and younger individuals seem to report psychosocial problems not identified by the HADS (58). Some of the other measures, e.g., the EROS, were chosen primarily to evaluate iCBT, but since that intervention barely took

77 place, their value can be questioned. The accuracy of using the PCL-C has previously been discussed, but it can be mentioned that changes were made in classification of post-traumatic stress in individuals with cancer in the DSM- V during the ongoing study. Some project-specific questions were used regarding perceived relevance, knowledge, and satisfaction. Since Step 1 involved assumptions from the SCT, it would have been valuable to add a measure of self-efficacy, e.g., to increase the possibilities to compare effects to those of other studies (208). A validated measure of self-efficacy in the cancer context that might have been appropriate is the Cancer Behaviour Inventory (209).

The conclusions in Paper II were drawn from the predefined, main analysis population (ITT population), thus including imputed data. In an attempt to validate the results, a sensitivity analysis was performed. This analysis par- tially supported the main analysis; however, no statistically significant differences were found, indicating some uncertainty. Also, statistical power was not maintained for the population of complete cases. The significance level of 5% means that we, if conducting 100 tests, accept that 5 of these are significant due to chance. Since the analysis included multiple testing, a limitation is the risk of false significances. Further, the reliability of the findings is limited by the fact that several participants were lost to follow-up. This is a well-known problem in internet-based studies often divided into non-usage attrition, where the intervention is not used, and drop-out attrition, where the assessments are not completed. Technology, programme features, as well as demographic and psychological characteristics of users, have been suggested to affect adherence (210). Drop-out may also occur when users are satisfied and have no further need for the intervention (156). These users are referred to as e-attainers or early completers (211), who may still benefit from the intervention, even if not completing measures (154). Since most visits (78%) in iCAN-DO took place within four months of randomisation, it may be hypothesised that the perceived usefulness decreased over time and thereby affected the response rate. It has been suggested that ITT analysis should be encouraged in internet- based interventions (154), supporting the drawing of conclusions from the ITT population in this study. The RCT design demands a rather static intervention content, but some suggestions to prolong the interest in future interventions could be the use of news feeds or more active forum structures (as suggested by informants in the qualitative studies). When validating intervention effects and attrition, it is an advantage if adherence is defined and assessed in a way that does not leave room for uncertainty. This is not a requirement when reg- istering trials at clinicaltrials.gov and is not an explicit statement in the Con- sort statement (203). However, it could be beneficial to consult a checklist, e.g., the TIDieR checklist for intervention description (212), and findings from previous research on adherence in internet-based trials, already in the planning stages of future trials.

78 Study 2 Adding the qualitative study alongside the RCT was important to further understanding of the results of Study 1. It would most likely have been even more valuable to conduct it earlier, as part of the initial acceptance test, to improve usefulness and ease of use before the start of the RCT. The purposeful sample strategy in Study 2 resulted in a sample of individuals with experiences from both steps of the intervention, which was important for the aim of the study. Still, the informants varied in sociodemographic measures, as well as in their general online activities. Meanwhile, the inclusion of individuals who used iCAN-DO to a low extent turned out to be a miscalculation. They had trouble expressing opinions about iCAN-DO and since it resulted in quite shallow interviews and most likely caused discomfort, the decision was made to conduct interviews with only two low-range users. However, it would have added value to interview them with a different research question, to further explore non-usage.

Since informants had access to a specific intervention, and this was the factor of interest, the anticipated specificity in their experiences may be considered to be high (213). Thus, 15 interviews provided enough material to achieve the aim of the study. The use of an interview guide contributed to the study by increasing stability and dependability (214, 215), but it was also a practical facilitator for the interviewer to let the informants talk freely, with the guide serving as a roadmap for the conversation. A limitation to be addressed as an aspect of confirmability is the fact that the interviewer worked within the project. Thus, informants may have expressed themselves more positively than if someone independent had conducted the interviews. However, an independent interviewer might have affected the quality in other ways, e.g., due to lack of knowledge about the features in iCAN-DO. Further, the interviewer was not convinced that the intervention would be seen as useful but was focused on reflecting the informants’ perceptions to enhance program development. The analysis was mainly conducted by the interviewer and then discussed and assessed by co-authors. This was possible since content analysis involves a systematic and tested process allowing replication of the procedure (216). Re- garding the transferability of results, the needs described by informants may be similar to other diagnostic groups. Hopefully the setting is described in a way that allows researchers, clinicians and others interested to determine if results are useful in other contexts. Finally, it should be mentioned that during the later interviews, an informant asked to be sent the transcript, which gave rise to the idea that all informants should have been offered this possibility. The approach is described in the literature as a part of a triangulation process to increase validity (217), and can be recommended for future studies.

79 Conclusions

 Internet-based stepped care improves self-reported symptoms of depression in individuals with newly diagnosed cancer compared with standard care. Low-intensity support seems to be more acceptable than iCBT in this setting. Including psycho-educative content regarding the diagnosis and treatment along with self-care advice seems appropriate in internet-based support targeting this group.

 Individuals with cancer experience several unmet needs in standard healthcare, leaving them with a lack of knowledge and emotional support relating to their cancer. Internet-based psychosocial support can serve as an important complement and future studies should be integrated with standard healthcare. Further, results suggest that evaluated internet-based support should be provided at the time of diagnosis and be available over a period of time to suit individual preferences.

 Co-creation to tailor an internet-based intervention will most likely increase the perceived usefulness of the intervention. It is highly important to conduct a thorough user analysis and strive for diversity in the input from those with lived experience. Including input from individuals from different backgrounds, of different ages, and with differing use of internet and social media, alongside the input from theories, will most likely add value to interventions.

 Usability of an internet-based intervention is affected by the perceived usefulness and ease of use of the intervention, as well as by the user’s circumstances in life and the consequences of the cancer and its treatment. Internet-based psychosocial support would gain from being possible to adapt based on individual preferences and being integrated into standard healthcare.

80 Clinical implications and further research

 Future efforts to improve the support to individuals with cancer are highly needed and future trials should focus on how to integrate and personalise low-intensity internet-based information and support as a natural part of clinical oncology care. Since the low-intensity support was not sufficient for all, further research is also needed on how to increase the efficacy and how to make iCBT feasible for this group. There is also a need for interventions targeting anxiety at different stages.

 In the design of future internet-based interventions, important stakehold- ers should be identified and involved. This includes both representatives from regular healthcare and those with lived experience of cancer. The research team should include all those with potential impact on the intervention, e.g., those designing the content and those designing the system. Best scientific evidence and theoretical perspectives, as well as aspects important when reporting the study, should be investigated and incorporated in the development.

 Future research is needed on how to define, assess and increase individuals’ adherence to internet-based interventions. Measures to prolong participants interest in future interventions could be the use of non-static content such as news feeds or more active forum structures, including health care staff in the discussions. Meanwhile, intention to treat analysis can be recommended as part of evaluations.

 The cost effectiveness of internet-based stepped care interventions needs to be examined further and there is also a need for further research regarding the most relevant PROMs in this setting. Screening is highly recommended in forthcoming interventions.

81 Svensk sammanfattning (Swedish summary)

Ett cancerbesked skapar ofta känslor av rädsla, nedstämdhet och oro. Många kan hantera den nya situationen med stöd av närstående och/eller sjukvårds- personal, medan andra får mer långvariga besvär med ångest och depression som följd. Tidigare studier har visat att personer med cancersjukdom inte får det psykosociala stöd de önskar och att det finns många hindrande faktorer när det gäller tillgången till stödjande vård. Många patienter använder sig av internet för att söka information i samband med cancerdiagnosen och internetbaserat stöd kan därför vara en naturlig utveckling av sedvanlig vård. Tidigare forskning inom området har visat att lågintensivt stöd innehållande stödjande omvårdnad, psykoedukation och egenvårdsråd kan ha positiva effekter avseende nedstämdhet och oro. Dock har tidigare studier i vissa fall varit bristfäl- liga gällande interventionernas beskrivning och design vilket gör resultaten svåra att tolka och värdera. Rekommendationer för fortsatt forskning har in- volverat tydligare beskrivningar av innehållet i interventioner, att stödet riktas till de personer som faktiskt har symtom enligt principen för stegvis vård (det vill säga att stödet ges till individen utifrån behov). Andra rekommendationer har varit att företrädare från sjukvården och personer med egen erfarenhet av cancer ska involveras i utvecklingen av interventionerna. Detta för att öka interventionens relevans och användbarhet, både i sedvanlig vård och för den tänkta målgruppen. Syftet med avhandlingen har varit att öka kunskaperna inom området genom att beskriva och undersöka design, effekter och upple- velser av ett internetbaserat, stegvis psykosocialt stöd för personer som nyli- gen diagnostiserats med en cancersjukdom och samtidigt rapporterar ångest- eller depressionssymtom.

Studie 1 Delarbete I I det första delarbetet beskrevs utvecklingen och innehållet av det första steget (steg 1; interaktivt stöd) i den internetbaserade stegvisa interventionen (iCAN- DO). Då tidigare forskning visat att psykoedukation i kombination med stöd till egenvård kan vara hjälpsamt så blev detta formen för steg 1. Inslag från relevanta teorier (exempelvis Orems egenvårdsteori) användes också i upp-

82 byggnaden. För att skapa ett innehåll med hög relevans för den tänkta mål- gruppen som också kunde fungera som ett komplement till sedvanlig vård, involverades kliniskt verksam personal och representanter från patientföre- ningar. Grupperna identifierade inom vilka områden kunskap och stöd behöv- des och utformade innehållet tillsammans med forskargruppen. Patientrepre- sentanterna uppmärksammade bland annat behovet av en intervention med tydlig koppling till cancersjukdomen, snarare än till de psykosociala proble- men, samt vikten av användbarhet i relation till de symtom som cancersjukdom och behandling kan innebära. Utvecklingen av steg 1 ledde till tre huvud- delar, ett bibliotek innehållande diagnosspecifika föreläsningar, samt psykoe- dukativ information och egenvårdsråd inom 16 huvudområden (som identifi- erades av arbetsgrupperna). Innehållet fanns tillgängligt i olika former så som texter, ljudfiler eller video. I steg 1 ingick även en modererad forumsektion där deltagarna kunde diskutera med varandra, och funktionen ”Fråga en Ex- pert” där deltagarna kunde ställa frågor om sjukdom och behandling till en sjuksköterska. Inom ramen för delarbete I presenterades slutligen data från ett acceptanstest där data avseende användning för de första 39 deltagarna i delarbete II extraherades och jämfördes med användardata från liknande internetbaserade interventioner. Då användningen av steg 1 visade sig motsvara dessa, så kunde den randomiserade kontrollerade studien (RCTn) som delarbete II avser fortgå utan ytterligare ändringar av interventionens innehåll.

Delarbete II I delarbete II undersöktes effekterna av den stegvisa interventionen (iCAN- DO) i en RCT. Deltagarna rekryterades till studien i samband med ett besök på sjukhuset. Vid medgivande fick de instruktioner om att logga in i U-CARE Portalen och där besvara screeninginstrumentet HADS (The Hospital Anxiety and Depression Scale). HADS består av två delskalor, HADS-A som mäter symtom på ångest och HADS-D som mäter symtom på depression. Samman- lagt screenades 909 deltagare och 27 % (n=245) rapporterade >7 (sannolika symtom) på någon av HADS delskalor. De 245 deltagarna randomiserades sedan till iCAN-DO-gruppen (n=124) eller till kontrollgruppen (n=121). Samtliga deltagare i iCAN-DO-gruppen hade tillgång till steg 1 under hela studietiden (10 mån). Majoriteten av deltagarna (85%) använde sig av materi- alet i steg 1, och de flesta besöken (78%) ägde rum inom 4 månader från randomiseringen. De deltagare som fortfarande hade symtom enligt någon av HADS delskalor (>7) vid 1 ,4 eller 7 månader fick erbjudande om steg 2; in- ternetbaserad kognitiv beteendeterapi (iKBT). Steg 2 erbjöds 82 av de 124 deltagarna, 16 personer accepterade och 7 av dessa slutförde minst en av mo- dulerna i steg 2. Vid det sista mättillfället (10 mån) hade iCAN-DO gruppens poäng på HADS-D (depression) minskat med 23 % och kontrollgruppens med 8% sedan första mättillfället. Analyserna visade en statistiskt signifikant skill- nad mellan grupperna avseende depression (p= <0.05). Skillnaderna avseende

83 HADS-A (ångest) var inte kliniskt eller statistiskt signifikanta. Gällande häl- sorelaterad livskvalitet sågs kliniska men inga statistiskt signifikanta skillna- der. Ett problem i delarbete II är att drygt hälften i iCAN-DO gruppen inte besvarade 10 månaders mätningen, vilket tillför en viss osäkerhet. Det här är ett välkänt problem i internetbaserade interventioner som kan ha flera möjliga förklaringar. Det kan delvis bero på faktorer relaterade till interventionen men också på att personer har en tendens att sluta besvara frågeformulär när de har fått ut det de tycker de behöver av en intervention. Att majoriteten av besöken i iCAN-DO ägde rum inom fyra månader från randomiseringen kan tala för detta.

Studie 2 Delarbete III-IV För att få en förståelse för hur deltagarna uppfattat användandet av iCAN-DO genomfördes intervjuer med 15 av deltagarna som hade haft tillgång till interventionen. Intervjuerna analyserades med hjälp av innehållsanalys. Delta- garna besvarade frågor rörande interventionens relevans och innehåll samt an- vändbarhet och tekniska förutsättningar. Studie 2 resulterade i delarbete III som handlar om innehållet och dess relevans, och delarbete IV som handlar om användbarheten.

Resultaten från delarbete III visade att deltagarna använde sig av det interak- tiva stödet i steg 1 som ett komplement till sedvanlig vård. Behovet av information beskrevs som mycket stort, framförallt i samband med diagnosbeskedet och det var ofta svårt att hinna få svar på sina frågor i samband med sjukhus- besöken. Deltagarna beskrev ofta iCAN-DO i relation till sedvanlig vård, där det kunde vara svårt att få kontakt, stöd och hjälp. Sökandet efter information på internet började tidigt, ibland före diagnosbeskedet, och det upplevdes som en trygghet att få tillgång till evidensbaserad information i steg 1. Att bli be- kräftad och få sina symtom förklarade beskrevs också som lugnande och hjälpsamt vilket antyder att utformningen av steg 1 var relevant. De deltagare som hade använt sig av både steg 1 och steg 2 (iKBT) beskrev hur de två stegen kompletterade varandra och det var också en fördel att både steg 1 och steg 2 kunde användas tillsammans med närstående. Det framträdde också att steg 1 verkade mer genomförbart för deltagarna än steg 2 under den pågående cancerbehandlingen. Deltagarna beskrev steg 2 som hjälpsamt gällande sym- tomkontroll, samtidigt som det faktum att programmet skulle genomföras under en utsatt tid medförde stress i en redan ansträngd livssituation. En del upplevde den internetbaserade formen av behandlingen och kontakten med psy- kologen som positiv medan andra beskrev den internetbaserade formen som

84 opersonlig och svår att få att fungera. Bland annat föreslogs det att ett person- ligt möte med hjälp av exempelvis video skulle fungera underlättande i början av behandlingen. Det hade också varit hjälpsamt för deltagarna om iCAN-DO hade integrerats ytterligare med sedvanlig vård. Resultaten från delarbete IV visade att symtomen av cancern och behandlingen påverkade deltagarnas an- vändning av iCAN-DO. Symtom så som trötthet och smärta gjorde det svårt att orka sitta vid datorn och fokusera på innehållet. Deltagarna beskrev det som underlättande att kunna ligga ner och att använda sig av surfplatta istället för dator och föreslog också att innehåll i exempelvis videofilmer skulle kunna delas upp i flera kortare delar. Tekniska problem beskrevs som något som på- verkade motivation och användning av interventionen. Det var värdefullt att kunna kontakta teknisk support. Innehållet i iCAN-DO beskrevs som viktigare än designen, men de som var erfarna datoranvändare med stort engagemang i sociala medier upplevde oftare designen som otillfredsställande jämfört med de som inte använde sig av sociala medier. Deltagarna uppskattade tillgäng- ligheten i iCAN-DO och beskrev många olika preferenser när det gällde att ta till sig innehållet, vilket antyder att de olika informationsformaten var en för- del.

Slutsatser  iCAN-DO hade positiva effekter på nedstämdhet/depression medan oro/ångest inte påverkades i lika stor utsträckning. Det lågintensiva stödet innehållande psykoedukation och egenvårdsstrategier verkade genomförbart för deltagarna medan iKBT upplevdes som mer an- strängande. Lågintensivt, internetbaserat stöd kan därför vara mest lämpligt i att introducera i sedvanlig vård. Eftersom många fortsätter att rapportera symtom behövs vidare studier om hur internetbaserat stöd kan effektiviseras, till exempel hur iKBT kan anpassas och göras mer lättillgängligt för den här gruppen.  Det är svårt för personer med cancer att få relevant information och stöd i sedvanlig vård. iCAN-DO fungerade som ett komplement eller istället för sedvanlig vård och upplevdes som hjälpsamt av deltagarna. Den upplevda användbarheten påverkades både av interventionen, av cancersjukdomen och av den egna livssituationen. Att samskapa interventioner med personer med egen erfarenhet av sjukdomen och representanter från sjukvården ökar sannolikt relevans och användbar- het. Vidare forskning behövs om hur internetbaserat stöd kan indivi- danpassas och införas i sedvanlig vård.

85 Acknowledgements

“Buy the ticket, take the ride” is an approach I like to embrace. I did not always keep to that path with ease, but I managed it – with the help of a whole lot of good people. My heartfelt thanks to everyone who has helped me. First of all, I would like to thank all the people with lived experience of cancer who have helped me and us during this project. Your generosity brings great hope. Also, many thanks to all the clinical experts who contributed to development, as well as all the healthcare staff enabling my work during enrolment in the clinical outpatient settings.

I am very grateful to my supervisors: First, and most importantly, Birgitta Jo- hansson, my head supervisor and one of the coolest, most sharp-witted, and dedicated people I know. I have learned incredibly from you. Thank you for challenging me and for believing that I was capable. My co-supervisors, Peter Nygren and Helena Igelström, have also contributed tremendously with their friendliness and willingness to share their expertise in both clinical and methodological matters. Thank you.

I would also like to thank Louise von Essen and all the members of the U- CARE team, for making my thesis possible and for including me in this context. Special thanks go to Ian Horne for his friendly helpfulness during data extraction and to frequent cowriter Sven Alfonsson, who besides improving my work, contributed with such a supportive and friendly attitude. Thank you.

The value of friendship should not be underestimated during the writing of a thesis and friendship is perhaps one of the best outcomes one can wish for. I am so thankful for the support from PhD student colleagues: Marina Forslund, whose kindness, warmth, rigidity, and support I couldn’t have managed without. Also, Anna Henriksson, who has been kind, supportive, and – not least – has helped me grasp the point of it all (1 Mos 3:19.). All members of the Life- style and Rehabilitation in Long-Term Illness Research Group also contributed with a friendly environment for input and support. Special thanks go to Cecilia Arving, Maria Carlsson, and Ingrid Anderzén, for your devoted work in lieu of my Kappa seminar. I am also very grateful for the support from my new friends at the Department of Clinical Research and Development (KFUE). Also, a special, heartfelt thank you to Eva Turup, for teaching me the true meaning of person-centred nursing at exactly the right time.

86 During my PhD studies, I have engaged in teaching and I would like to thank all my students, co-workers, and the administrative staff, past and present, at the Department of Surgical Sciences. A very special thanks to my colleague Eva Kvidal, who has served as a brilliant role model. Our cooperation has enriched me, and I am so thankful for the opportunity to learn from someone so experienced. Also, many thanks to my colleague Johan Lingsarve, for often helping me out with such calm and friendly attitude.

Fortunately, there is more to life than work and I am most grateful for the support from family and friends. First of all, mum and dad: When I went through my phonebook to make sure I didn’t forget to thank anyone important, I realised that you are still listed there as “Home, sweet home”, even though I have had my own home for so many years. I don’t have the heart to change it. Thank you.

I am grateful for having my brother Erik and sister Kerstin with families close by. Thank you for always supporting me. Special thanks go to the smart and caring Vide and Ester – thank you for being such good mates to me and your cousin! Also, a special welcome to little Greta, who I look forward to getting to know. In addition, I am lucky to have a bunch of caring relatives, cousins, and aunts of my own, most importantly: Inga, Gerd, Anna, Gunnar, Inger, and Christer, who have always been so kind to me. Special thanks to the Olsson kids for making my summers active and playful!

I would also like to thank Magdalena Eldelind, for being a great friend and for sharing a positive outlook on life that has been vital for me over the years. Many thanks to Agnes Dunder, John Flemström, and Knut Almgren, who always make me feel included in their hearts of steel as I fight the world, and to Wilhelmina Allard for making me a better person by sharing her empathetic, idealistic mind. If you don’t give up on me, I won’t give up on you. Thank you.

Most importantly, I am so grateful to you, Ola and Ivar. Sometimes I can’t believe all that we have managed together. Your unconditional love always tips me back into balance when I stumble. I would be truly lost without you. Thank you for putting up with my stress and adjusting your lives for me to be able to wrap this up. Now, the lovely summer months await, and I so look forward to sharing them with you. Jag älskar er så.

87 References

1. Farre A, Rapley T. The New Old (and Old New) Medical Model: Four Decades Navigating the Biomedical and Psychosocial Understandings of Health and Illness. Healthcare (Basel, Switzerland). 2017;5(4). 2. Caplan AL, McCartney JJ, Sisti DA. Health, disease, and illness: concepts in medicine. Washington, D.C: Georgetown University Press; 2004. 3. Antonovsky A, Elfstadius M. Hälsans mysterium. Stockholm: Natur och kultur; 2005. 4. Engel GL. The Need for a New Medical Model: A Challenge for Biomedicine. Holistic Medicine. 1989;4(1):37-53. 5. World Health Organization. WHO definition of health [cited 2020 -04-16]. [Available from: https://www.who.int/about/who-we-are/constitution. 6. Alligood MR. Nursing theorists and their work. 9 ed. St. Louis, Missouri: Else- vier; 2017. 7. The Swedish Society of Nursing. Kompetensbeskrivning för legitimerad Sjuk- sköterska [cited 2020 -04-16]. Available from: https://www.swenurse.se/Sa- tycker-vi/publikationer/Kompetensbeskrivningar-och-riktlinjer/kompetensbeskrivning-for-legitimerad-sjukskoterska/. 8. Mead N, Bower P. Patient-centredness: a conceptual framework and review of the empirical literature. Social Science & Medicine. 2000;51(7):1087-110. 9. Singer S. Psychosocial Impact of Cancer. Psychooncology Recent Results in Cancer Research: Springer, Cham; 2018. p. 1-11. 10. Cohen MZ. The meaning of cancer and oncology nursing: link to effective care. Seminars in oncology nursing. 1995;11(1):59-67. 11. Moghaddam N, Coxon H, Nabarro S, Hardy B, Cox K. Unmet care needs in people living with advanced cancer: a systematic review. Supportive Care in Cancer. 2016;24(8):3609-22. 12. Cheville AL, Mustian K, Winters-Stone K, Zucker DS, Gamble GL, Alfano CM. Cancer Rehabilitation: An Overview of Current Need, Delivery Models, and Levels of Care. Physical medicine and rehabilitation clinics of North America. 2017;28(1):1-17. 13. Ebel M-D, Stellamanns J, Keinki C, Rudolph I, Huebner J. Cancer Patients and the Internet: a Survey Among German Cancer Patients. Journal of Cancer Ed- ucation. 2017;32(3):503-8. 14. Finney Rutten LJ, Agunwamba AA, Wilson P, Chawla N, Vieux S, Blanch- Hartigan D, et al. Cancer-Related Information Seeking Among Cancer Survi- vors: Trends Over a Decade (2003–2013). Journal of Cancer Education. 2016;31(2):348-57. 15. The Swedish Cancer Society. Cancerfondsrapporten 2018 [cited 2020 -04-16]. Available from: https://static-files.cancerfonden.se/Cancerfondsrappor- ten2018_webb_(2)_1521607903.pdf. 16. World Health Organization. Cancer Key facts [cited 2020 -04-16]. Available from: https://www.who.int/news-room/fact-sheets/detail/cancer.

88 17. Frojd C, Larsson G, Lampic C, von Essen L. Health related quality of life and psychosocial function among patients with carcinoid tumours. A longitudinal, prospective, and comparative study. Health and Quality of Life Outcomes. 2007;5(1):18. 18. Maass SW, Roorda C, Berendsen AJ, Verhaak PF, de Bock GH. The prevalence of long-term symptoms of depression and anxiety after breast cancer treatment: A systematic review. Maturitas. 2015;82(1):100-8. 19. Mitchell AJ, Ferguson DW, Gill J, Paul J, Symonds P. Depression and anxiety in long-term cancer survivors compared with spouses and healthy controls: a systematic review and meta-analysis. The Lancet Oncology. 2013;14(8):721- 32. 20. Dunkel-Schetter C, Feinstein LG, Taylor SE, Falke RL. Patterns of coping with cancer. Health psychology : official journal of the Division of Health Psychol- ogy, American Psychological Association. 1992;11(2):79-87. 21. Cao W, Qi X, Yao T, Han X, Feng X. How doctors communicate the initial diagnosis of cancer matters: cancer disclosure and its relationship with Pa- tients' hope and trust. Psycho-Oncology. 2017;26(5):640-8. 22. Nekanda-Trepka CJ, Bishop S, Blackburn IM. Hopelessness and depression. Br J Clin Psychol. 1983;22 Pt 1:49-60. 23. Rivera DR, Ganz PA, Weyrich MS, Bandos H, Melnikow J. Chemotherapy- Associated Peripheral Neuropathy in Patients With Early-Stage Breast Cancer: A Systematic Review. J Natl Cancer Inst. 2018;110(2). 24. Browne BM, Vanni AJ. Management of Urethral Stricture and Bladder Neck Contracture Following Primary and Salvage Treatment of Prostate Cancer. Current Urology Reports. 2017;18(10):76. 25. Cullberg J. Kris och utveckling : en psykodynamisk och socialpsykiatrisk studie. Stockholm: Natur och kultur; 2003. 26. Cella DF, Tulsky DS. Quality of life in cancer: definition, purpose, and method of measurement. Cancer investigation. 1993;11(3):327-36. 27. Djarv T, Wikman A, Johar A, Lagergren P. Poor health-related quality of life in the Swedish general population: the association with disease and lifestyle factors. Scandinavian journal of public health. 2013;41(7):744-53. 28. Reilly CM, Bruner DW, Mitchell SA, Minasian LM, Basch E, Dueck AC, et al. A literature synthesis of symptom prevalence and severity in persons receiving active cancer treatment. Supportive Care in Cancer. 2013;21(6):1525- 50. 29. Barsevick AM, Whitmer K, Nail LM, Beck SL, Dudley WN. Symptom Cluster Research: Conceptual, Design, Measurement, and Analysis Issues. Journal of Pain and Symptom Management. 2006;31(1):85-95. 30. Beck SL, Dudley WN, Barsevick A. Pain, sleep disturbance, and fatigue in patients with cancer: using a mediation model to test a symptom cluster. On- cology nursing forum. 2005;32(3):542. 31. Knowles G, Haigh R, McLean C, Phillips H. Late effects and quality of life after chemo-radiation for the treatment of anal cancer. European journal of oncology nursing : the official journal of European Oncology Nursing Society. 2015;19(5):479-85. 32. Ness KK, Wall MM, Oakes JM, Robison LL, Gurney JG. Physical Perfor- mance Limitations and Participation Restrictions Among Cancer Survivors: A Population-Based Study. Annals of Epidemiology.16(3):197-205. 33. Williamson GM, Schulz R. Activity restriction mediates the association between pain and depressed affect: a study of younger and older adult cancer patients. Psychology and aging. 1995;10(3):369-78.

89 34. Sjovall K, Attner B, Lithman T, Noreen D, Gunnars B, Thome B, et al. Influ- ence on the health of the partner affected by tumor disease in the wife or hus- band based on a population-based register study of cancer in Sweden. J Clin Oncol. 2009;27(28):4781-6. 35. Blindheim K, Thorsnes SL, Brataas HV, Dahl BM. The role of next of kin of patients with cancer: learning to navigate unpredictable caregiving situations. Journal of Clinical Nursing. 2013;22(5-6):681-9. 36. Swedish municipalities and regions in cooperation. National patient inquiry- standardized care course in cancer care [cited 2020 -02-03]. Available from: https://patientenkat.se/sv/resultat/Standardiserade%20vårdför- lopp%20inom%20cancervård%202019. 37. Cheville AL, Kornblith AB, Basford JR. An examination of the causes for the underutilization of rehabilitation services among people with advanced cancer. American journal of physical medicine & rehabilitation. 2011;90(5 Suppl 1):S27-37. 38. Thalen-Lindstrom A, Larsson G, Glimelius G, Johansson B. Anxiety and depression in oncology patients; a longitudinal study of a screening, assessment and psychosocial support intervention. Acta Oncologica Jan 2013. 2013;52(1):118 - 27. 39. Wiklund Gustin L. Psykologi för sjuksköterskor. Lund: Studentlitteratur; 2015. 40. Yi JC, Syrjala KL. Anxiety and Depression in Cancer Survivors. The Medical clinics of North America. 2017;101(6):1099-113. 41. Bandura A. Self-efficacy: Toward a unifying theory of behavioral change. Psy- chological Review. 1977;84(2):191-215. 42. Zubin J, Spring B. Vulnerability--a new view of schizophrenia. Journal of ab- normal psychology. 1977;86(2):103-26. 43. Morrison AP. A casebook of cognitive therapy for psychosis. Hove: Brunner- Routledge; 2002. p. 59-73. 44. Mehnert A, Hartung TJ, Friedrich M, Vehling S, Brähler E, Härter M, et al. One in two cancer patients is significantly distressed: Prevalence and indicators of distress. Psycho-Oncology. 2018;27(1):75-82. 45. Mitchell AJ, Chan M, Bhatti H, Halton M, Grassi L, Johansen C, et al. Preva- lence of depression, anxiety, and adjustment disorder in oncological, haematological, and palliative-care settings: a meta-analysis of 94 interview-based studies. The Lancet Oncology. 2011;12(2):160-74. 46. Thalen-Lindstrom A, Glimelius B, Johansson B. Development of anxiety, depression and health-related quality of life in oncology patients without initial symptoms according to the Hospital Anxiety and Depression Scale - a comparative study. Acta oncologica. 2017;56(8):1094-102. 47. Pitman A, Suleman S, Hyde N, Hodgkiss A. Depression and anxiety in patients with cancer. Bmj. 2018;361:k1415. 48. Smith HR. Depression in cancer patients: Pathogenesis, implications and treatment (Review). Oncol Lett. 2015;9(4):1509-14. 49. Braamse AM, van Turenhout ST, Terhaar Sive Droste JS, de Groot GH, van der Hulst RW, Klemt-Kropp M, et al. Factors associated with anxiety and depressive symptoms in colorectal cancer survivors. European journal of gastro- enterology & hepatology. 2016;28(7):831-5. 50. Vehling S, Kissane DW. Existential distress in cancer: Alleviating suffering from fundamental loss and change. Psycho-Oncology. 2018;27(11):2525-30. 51. Ridner SH. Psychological distress: concept analysis. Journal of advanced nursing. 2004;45(5):536-45.

90 52. Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale. Acta Psychiat Scand. 1983;67(6):361-70. 53. Ibbotson T, Maguire P, Selby P, Priestman T, Wallace L. Screening for anxiety and depression in cancer patients: the effects of disease and treatment. Eur J Cancer. 1994;30a(1):37-40. 54. Brintzenhofe-Szoc KM, Levin TT, Li Y, Kissane DW, Zabora JR. Mixed anxiety/depression symptoms in a large cancer cohort: prevalence by cancer type. Psychosomatics. 2009;50(4):383-91. 55. Hinz A, Herzberg PY, Lordick F, Weis J, Faller H, Brähler E, et al. Age and gender differences in anxiety and depression in cancer patients compared with the general population. European journal of cancer care. 2019;28(5):e13129. 56. Linden W, Vodermaier A, MacKenzie R, Greig D. Anxiety and depression after cancer diagnosis: Prevalence rates by cancer type, gender, and age. Jour- nal of Affective Disorders. 2012;141(2–3):343-51. 57. Rice SM, Kealy D, Ogrodniczuk JS, Seidler ZE, Denehy L, Oliffe JL. The Cost of Bottling It Up: Emotion Suppression as a Mediator in the Relationship Be- tween Anger and Depression Among Men with Prostate Cancer. Cancer management and research. 2020;12:1039-46. 58. Thalen-Lindstrom AM, Glimelius BG, Johansson BB. Identification of Dis- tress in Oncology Patients: A Comparison of the Hospital Anxiety and Depres- sion Scale and a Thorough Clinical Assessment. Cancer Nurs. 2016;39(2):E31- 9. 59. Watts S, Leydon G, Birch B, Prescott P, Lai L, Eardley S, et al. Depression and anxiety in prostate cancer: a systematic review and meta-analysis of prevalence rates. BMJ open. 2014;4(3):e003901. 60. Krebber AMH, Buffart LM, Kleijn G, Riepma IC, de Bree R, Leemans CR, et al. Prevalence of depression in cancer patients: a meta-analysis of diagnostic interviews and self-report instruments. Psycho-Oncology. 2014;23(2):121-30. 61. Dalton SO, Laursen TM, Ross L, Mortensen PB, Johansen C. Risk for Hospi- talization With Depression After a Cancer Diagnosis: A Nationwide, Popula- tion-Based Study of Cancer Patients in Denmark From 1973 to 2003. Journal of Clinical Oncology. 2009;27(9):1440-5. 62. Pinquart M, Duberstein P. Depression and cancer mortality: a meta-analysis. Psychol Med. 2010;40(11):1797 - 810. 63. Robson A, Scrutton F, Wilkinson L, MacLeod F. The risk of suicide in cancer patients: a review of the literature. Psycho-Oncology. 2010;19(12):1250-8. 64. Carlsson S, Sandin F, Fall K, Lambe M, Adolfsson J, Stattin P, et al. Risk of suicide in men with low-risk prostate cancer. European Journal of Cancer. 2013;49(7):1588-99. 65. Mausbach BT, Yeung P, Bos T, Irwin SA. Health care costs of depression in patients diagnosed with cancer. Psychooncology. 2018;27(7):1735-41. 66. DiMatteo M, Lepper HS, Croghan TW. Depression is a risk factor for noncom- pliance with medical treatment: Meta-analysis of the effects of anxiety and depression on patient adherence. Archives of Internal Medicine. 2000;160(14):2101-7. 67. Arrieta Ó, Angulo LP, Núñez-Valencia C, Dorantes-Gallareta Y, Macedo EO, Martínez-López D, et al. Association of Depression and Anxiety on Quality of Life, Treatment Adherence, and Prognosis in Patients with Advanced Non- small Cell Lung Cancer. Annals of Surgical Oncology. 2013;20(6):1941-8. 68. Sollner W, DeVries A, Steixner E, Lukas P, Sprinzl G, Rumpold G, et al. How successful are oncologists in identifying patient distress, perceived social support, and need for psychosocial counselling? Br J Cancer. 2001;84(2):179-85.

91 69. Hedstrom M, Kreuger A, Ljungman G, Nygren P, von Essen L. Accuracy of assessment of distress, anxiety, and depression by physicians and nurses in adolescents recently diagnosed with cancer. Pediatric blood & cancer. 2006;46(7):773-9. 70. Fallowfield L, Ratcliffe D, Jenkins V, Saul J. Psychiatric morbidity and its recognition by doctors in patients with cancer. Brit J Cancer. 2001;84:1011. 71. Mitchell AJ, Kaar S, Coggan C, Herdman J. Acceptability of common screening methods used to detect distress and related mood disorders-preferences of cancer specialists and non-specialists. Psychooncology. 2008;17(3):226-36. 72. McHorney CA. Generic health measurement: past accomplishments and a measurement paradigm for the 21st century. Annals of internal medicine. 1997;127(8 Pt 2):743-50. 73. Howell D, Fitch M, Bakker D, Green E, Sussman J, Mayo S, et al. Core domains for a person-focused outcome measurement system in cancer (PROMS- Cancer Core) for routine care: a scoping review and Canadian Delphi Consen- sus. Value in health : the journal of the International Society for Pharmacoeco- nomics and Outcomes Research. 2013;16(1):76-87. 74. Nelson EC, Eftimovska E, Lind C, Hager A, Wasson JH, Lindblad S. Patient reported outcome measures in practice. BMJ : British Medical Journal. 2015;350:g7818. 75. Mårdberg B, Carlstedt L. Grundläggande psykometri. Lund: Studentlitteratur; 2019. 76. Frost MH, Reeve BB, Liepa AM, Stauffer JW, Hays RD. What is sufficient evidence for the reliability and validity of patient-reported outcome measures? Value in health : the journal of the International Society for Pharmacoeconom- ics and Outcomes Research. 2007;10 Suppl 2:S94-s105. 77. Bring J, Taube A, Wikman P. Introduktion till medicinsk statistik. Lund: Stu- dentlitteratur; 2015. 78. McGlothlin AE, Lewis RJ. Minimal clinically important difference: Defining what really matters to patients. JAMA. 2014;312(13):1342-3. 79. Schünemann HJ, Guyatt GH. Commentary--goodbye M(C)ID! Hello MID, where do you come from? Health services research. 2005;40(2):593-7. 80. Revicki D, Hays RD, Cella D, Sloan J. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. Journal of clinical epidemiology. 2008;61(2):102-9. 81. Lewin S, Glenton C, Oxman AD. Use of qualitative methods alongside randomised controlled trials of complex healthcare interventions: methodological study. Bmj. 2009;339. 82. Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, et al. The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst. 1993;85(5):365-76. 83. Cella DF, Tulsky DS, Gray G, Sarafian B, Linn E, Bonomi A, et al. The Func- tional Assessment of Cancer Therapy scale: development and validation of the general measure. J Clin Oncol. 1993;11(3):570-9. 84. Luckett T, King MT, Butow PN, Oguchi M, Rankin N, Price MA, et al. Choos- ing between the EORTC QLQ-C30 and FACT-G for measuring health-related quality of life in cancer clinical research: issues, evidence and recommendations. Annals of Oncology. 2011;22(10):2179-90. 85. Aminisani N, Nikbakht H, Asghari Jafarabadi M, Shamshirgaran SM. Depres- sion, anxiety, and health related quality of life among colorectal cancer survivors. Journal of gastrointestinal oncology. 2017;8(1):81-8.

92 86. Donovan KA, Jacobsen PB. Fatigue, Depression, and Insomnia: Evidence for a Symptom Cluster in Cancer. Seminars in oncology nursing. 2007;23(2):127- 35. 87. Morin CM, Belleville G, Bélanger L, Ivers H. The Insomnia Severity Index: Psychometric Indicators to Detect Insomnia Cases and Evaluate Treatment Re- sponse. Sleep. 2011;34(5):601-8. 88. Yusufov M, Zhou ES, Recklitis CJ. Psychometric properties of the Insomnia Severity Index in cancer survivors. Psychooncology. 2019;28(3):540-6. 89. Yellen SB, Cella DF, Webster K, Blendowski C, Kaplan E. Measuring fatigue and other anemia-related symptoms with the Functional Assessment of Cancer Therapy (FACT) measurement system. Journal of Pain and Symptom Manage- ment. 1997;13(2):63-74. 90. Mitchell AJ, Meader N, Symonds P. Diagnostic validity of the Hospital Anxi- ety and Depression Scale (HADS) in cancer and palliative settings: a meta- analysis. J Affect Disord. 2010;126(3):335-48. 91. Bjelland I, Dahl A, Haug T, Neckelmann D. The validity of the Hospital Anx- iety and Depression Scale: An updated literature review. J Psychosom Res. 2002;52:69 - 77. 92. Kvaal K, Ulstein I, Nordhus I, Engedal K. The Spielberger statetrait anxiety inventory (STAI): the state scale in detecting mental disorders in geriatric patients. Int J Geriatric Psychiatry. 2005;20(7):629 - 34. 93. Marziliano A, Tuman M, Moyer A. The relationship between post-traumatic stress and post-traumatic growth in cancer patients and survivors: A systematic review and meta-analysis. Psycho-Oncology. 2019;n/a(n/a). 94. U.S Department of Veteran Affairs. PTSD Checklist for DSM-5 (PCL-5) [cited 20 04-16]. Available from: https://www.ptsd.va.gov/professional/assessment/adult-sr/ptsd-checklist.asp#obtain. 95. Horowitz M, Wilner N, Alvarez W. Impact of Event Scale: a measure of subjective stress. Psychosom Med. 1979;41(3):209-18. 96. Priede A, Andreu Y, Martínez P, Conchado A, Ruiz-Torres M, González- Blanch C. The factor structure of the Medical Outcomes Study–Social Support Survey: A comparison of different models in a sample of recently diagnosed cancer patients. Journal of Psychosomatic Research. 2018;108:32-8. 97. Derogatis LR. The psychosocial adjustment to illness scale (PAIS). Journal of Psychosomatic Research. 1986;30(1):77-91. 98. Woodward K. Psychosocial studies. Hoboken: Taylor and Francis; 2015. p. 1- 14. 99. Galway K, Black A, Cantwell M, Cardwell CR, Mills M, Donnelly M. Psy- chosocial interventions to improve quality of life and emotional wellbeing for recently diagnosed cancer patients. Cochrane Db Syst Rev. 2012;11:Cd007064. 100. Sheard T, Maguire P. The effect of psychological interventions on anxiety and depression in cancer patients: results of two meta-analyses. Brit J Cancer. 1999;80(11):1770-80. 101. Faller H, Schuler M, Richard M, Heckl U, Weis J, Kuffner R. Effects of psycho-oncologic interventions on emotional distress and quality of life in adult patients with cancer: systematic review and meta-analysis. J Clin Oncol. 2013;31(6):782-93. 102. Sanjida S, McPhail SM, Shaw J, Couper J, Kissane D, Price MA, et al. Are psychological interventions effective on anxiety in cancer patients? A systematic review and meta-analyses. Psychooncology. 2018(9):2063-76.

93 103. Swedish National Board of Health and Welfare. The development of Regional Cancer Centres [cited 20 -03-19]. Available from: https://www.socialstyrelsen.se/globalassets/sharepoint-dokument/artikelkatalog/ovrigt/2017-10- 36.pdf. 104. Swedish Regional Cancer Centers in Cooperation. Tydligare uppdrag för kon- taktsjuksköterskan [cited 20 -01-31]. Available from: https://www.cancercentrum.se/samverkan/om-oss/nyheter/2019/november/kontaktsjukskoter- skans-uppdrag-fortydligas/. 105. Söderlund M. "En mänsklig atmosfär" : trygghet, samhörighet och gemenskap : god vård ur ett patientperspektiv = ["In the atmosphere of humanity"] : [security, solidarity and mutual understanding] : [a concept of good caring from the patients' perspective]. Uppsala: Acta Universitatis Upsaliensis :; 1998. 106. Swedish Ministry of Health and Social Affairs. En nationell cancerstrategi för framtiden [cited 20 -04-16]. Available from: http://www.regeringen.se/rattsdokument/statens-offentliga-utred- ningar/2009/02/sou-200911/. 107. The Swedish Cancer Society. Om Cancer [cited 20 04-16]. Available from: https://www.cancerfonden.se/om-cancer/om-cancer. 108. Lawler M, Banks I, Law K, Albreht T, Armand J-P, Barbacid M, et al. The European Cancer Patient’s Bill of Rights, update and implementation 2016. ESMO Open. 2016;1(6). 109. Dilworth S, Higgins I, Parker V, Kelly B, Turner J. Patient and health professional's perceived barriers to the delivery of psychosocial care to adults with cancer: a systematic review. Psycho-Oncology. 2014;23(6):601-12. 110. Ohlsson-Nevo E, Alkebro I, Ahlgren J. Cancer patients' interest in participating in cancer rehabilitation. Acta oncologica. 2019;58(12):1676-83. 111. Linton SJ, Flink I. 12 verktyg i KBT : från teori till färdighet. Stockholm: Natur & kultur; 2011. p. 55-65. 112. American Psychological Association. What is Cognitive Behavioral Therapy? [cited 20 -03-25]. Available from: https://www.apa.org/ptsd-guideline/patients-and-families/cognitive-behavioral. 113. McCorkle R, Ercolano E, Lazenby M, Schulman-Green D, Schilling LS, Lorig K, et al. Self-management: Enabling and empowering patients living with cancer as a chronic illness. CA Cancer J Clin. 2011;61(1):50-62. 114. Cuthbert CA, Farragher JF, Hemmelgarn BR, Ding Q, McKinnon GP, Cheung WY. Self-management interventions for cancer survivors: A systematic review and evaluation of intervention content and theories. Psycho-Oncology. 2019;28(11):2119-40. 115. Kim SH, Kim K, Mayer DK. Self-Management Intervention for Adult Cancer Survivors After Treatment: A Systematic Review and Meta-Analysis. Oncol- ogy nursing forum. 2017;44(6):719-28. 116. Linden W, Girgis A. Psychological treatment outcomes for cancer patients: what do meta-analyses tell us about distress reduction? Psychooncology. 2012;21(4):343-50. 117. Phd F, Ncama B. Self, self-care and self-management concepts: Implications for self-management education. Education Research. 2011;2:1733-7. 118. Orem DE, Taylor SG, Renpenning KM. Nursing : concepts of practice. St. Louis, Mo.: Mosby; 2001. 119. Wilson J, Gramling L. The application of Orem's Self-Care Model to burn care. Journal of burn care & research : official publication of the American Burn Association. 2009;30(5):852-8.

94 120. Bernier F. Relationship of a pelvic floor rehabilitation program for urinary in- continence to Orem's Self-Care Deficit Theory of Nursing: Part 1. Urologic nursing. 2002;22(6):378-83, 90; quiz 91. 121. Baker LK, Denyes MJ. Predictors of Self-Care in Adolescents With Cystic Fi- brosis: A Test of Orem's Theories of Self-Care and Self-Care Deficit. Journal of Pediatric Nursing. 2008;23(1):37-48. 122. Jaarsma T, Halfens R, Senten M, Abu Saad HH, Dracup K. Developing a supportive-educative program for patients with advanced heart failure within Orem's general theory of nursing. Nursing science quarterly. 1998;11(2):79- 85. 123. Dodd MJ, Dibble SL. Predictors of self-care: a test of Orem's model. Oncology nursing forum. 1993;20(6):895-901. 124. Desbiens J-F, Gagnon J, Fillion L. Development of a shared theory in palliative care to enhance nursing competence. Journal of advanced nursing. 2012;68(9):2113-24. 125. Richard AA, Shea K. Delineation of Self-Care and Associated Concepts. Jour- nal of Nursing Scholarship. 2011;43(3):255-64. 126. The Swedish Agency for Health Technology Assessment and Assessment of Social Services. Behandling av depressionssjukdomar. [cited 20 -02-18]. Available from: https://www.sbu.se/sv/publikationer/SBU-utvarderar/behandling-av-depressionssjukdomar/. 127. Arico D, Raggi A, Ferri R. Cognitive Behavioral Therapy for Insomnia in Breast Cancer Survivors: A Review of the Literature. Front Psychol. 2016;7:1162. 128. Johnson JA, Rash JA, Campbell TS, Savard J, Gehrman PR, Perlis M, et al. A systematic review and meta-analysis of randomized controlled trials of cognitive behavior therapy for insomnia (CBT-I) in cancer survivors. Sleep Medi- cine Reviews. 2016;27:20-8. 129. The Swedish Agency for Health Technology Assessment and Assessment of Social Services. Bröstcancer och rehabilitering [20-02-18]. Available from: https://www.sbu.se/brostcancer_rehab. 130. Serfaty M, King M, Nazareth I, Moorey S, Aspden T, Mannix K, et al. Effec- tiveness of cognitive-behavioural therapy for depression in advanced cancer: CanTalk randomised controlled trial. Br J Psychiatry. 2019:1-9. 131. Moorey S, Greer S. Oxford guide to CBT for people with cancer Oxford: Ox- ford University Press; 2011. p. 61-76. 132. Bower P, Gilbody S. Stepped care in psychological therapies: access, effectiveness and efficiency. Narrative literature review. Br J Psychiatry. 2005;186:11 - 7. 133. Tamminga SJ, Hoving JL, Frings-Dresen MH, de Boer AG. Cancer@Work - a nurse-led, stepped-care, e-health intervention to enhance the return to work of patients with cancer: study protocol for a randomized controlled trial. Trials. 2016;17(1):453. 134. ITU. Measuring the information society report. [18-04-12]. Available from: https://www.itu.int/en/ITU-D/Statistics/Documents/publications/misr2017/MISR2017_Volume1.pdf. 135. The Swedish Internet Foundation. The Swedes and the Internet. [20-04-16]. Available from: https://svenskarnaochinternet.se/rapporter/svenskarna-och-internet-2019/sammanfattning/. 136. Mattsson S, Olsson EMG, Johansson B, Carlsson M. Health-Related Internet Use in People With Cancer: Results From a Cross-Sectional Study in Two Outpatient Clinics in Sweden. J Med Internet Res. 2017;19(5):e163.

95 137. Swedish Regional Cancer Centers in Cooperation. Min vårdplan i Stöd och behandling-Utvärdering av piloter [cited 18-04-10. Available from: https://www.cancercentrum.se/globalassets/patient-och-narstaende/samverkan/utvardering-pilot-min-vardplan-cancer-2017.pdf. 138. Oh H, Rizo C, Enkin M, Jadad A. What Is eHealth (3): A Systematic Review of Published Definitions. J Med Internet Res. 2005;7(1):e1. 139. Swedish National Board of Health and Welfare. E-hälsa. [20-04-16]. Availa- ble from: https://www.socialstyrelsen.se/utveckla-verksamhet/e-halsa/. 140. The Swedish Society of Nursing. Strategi för sjuksköterskors arbete med e- hälsa. [20-04-16]. Available from: https://www.swenurse.se/Sa-tycker- vi/publikationer/eHalsa/Strategi-for-sjukskoterskors-arbete-med-eHalsa/. 141. Murray E, Burns J, See T, Lai R, Nazareth I. Interactive health communication applications for people with chronic disease. Cochrane Database Syst Rev. 2005;4:12 - 23. 142. Bouma G, Admiraal JM, de Vries EGE, Schröder CP, Walenkamp AME, Reyners AKL. Internet-based support programs to alleviate psychosocial and physical symptoms in cancer patients: A literature analysis. Critical Reviews in Oncology/Hematology. 2015;95(1):26-37. 143. Chalmers I, Glasziou P. Avoidable waste in the production and reporting of research evidence. Lancet. 2009;374(9683):86-9. 144. Ruland CM, Andersen T, Jeneson A, Moore S, Grimsbo GH, Borosund E, et al. Effects of an internet support system to assist cancer patients in reducing symptom distress: a randomized controlled trial. Cancer Nurs. 2013;36(1):6- 17. 145. Ruland CM, Jeneson A, Andersen T, Andersen R, Slaughter L, Bente Schjodt O, et al. Designing tailored Internet support to assist cancer patients in illness management. AMIA Annu Symp Proc. 2007:635-9. 146. McTavish FM, Gustafson DH, Owens BH, Hawkins RP, Pingree S, Wise M, et al. CHESS (Comprehensive Health Enhancement Support System): an interactive computer system for women with breast cancer piloted with an underserved population. The Journal of ambulatory care management. 1995;18(3):35-41. 147. Gustafson DH, Hawkins R, Pingree S, McTavish F, Arora NK, Mendenhall J, et al. Effect of computer support on younger women with breast cancer. Jour- nal of general internal medicine. 2001;16(7):435-45. 148. Hedman E, Ljótsson B, Lindefors N. Cognitive behavior therapy via the Inter- net: a systematic review of applications, clinical efficacy and cost–effectiveness. Expert Review of Pharmacoeconomics & Outcomes Research. 2012;12(6):745-64. 149. Andersson G, Cuijpers P, Carlbring P, Riper H, Hedman E. Guided Internet- based vs. face-to-face cognitive behavior therapy for psychiatric and somatic disorders: a systematic review and meta-analysis. World psychiatry : official journal of the World Psychiatric Association (WPA). 2014;13(3):288-95. 150. Mehta S, Peynenburg VA, Hadjistavropoulos HD. Internet-delivered cognitive behaviour therapy for chronic health conditions: a systematic review and meta- analysis. J Behav Med. 2019;42(2):169-87. 151. Abrahams HJG, Gielissen MFM, Donders RRT, Goedendorp MM, van der Wouw AJ, Verhagen CAHHVM, et al. The efficacy of Internet-based cognitive behavioral therapy for severely fatigued survivors of breast cancer compared with care as usual: A randomized controlled trial. Cancer. 2017;123(19):3825-34.

96 152. Compen F, Bisseling E, Schellekens M, Donders R, Carlson L, van der Lee M, et al. Face-to-Face and Internet-Based Mindfulness-Based Cognitive Therapy Compared With Treatment as Usual in Reducing Psychological Distress in Pa- tients With Cancer: A Multicenter Randomized Controlled Trial. Journal of Clinical Oncology. 2018;36(23):2413-21. 153. Murphy MJ, Newby JM, Butow P, Loughnan SA, Joubert AE, Kirsten L, et al. Randomised controlled trial of internet-delivered cognitive behaviour therapy for clinical depression and/or anxiety in cancer survivors (iCanADAPT Early). Psychooncology. 2019. 154. Christensen H, Griffiths KM, Farrer L. Adherence in Internet Interventions for Anxiety and Depression: Systematic Review. J Med Internet Res. 2009;11(2):e13. 155. Sieverink F, Kelders SM, van Gemert-Pijnen JE. Clarifying the Concept of Adherence to eHealth Technology: Systematic Review on When Usage Be- comes Adherence. J Med Internet Res. 2017;19(12):e402-e. 156. Eysenbach G. The law of attrition. J Med Internet Res. 2005;7(1):e11. 157. Davis FD. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly. 1989;13(3):319-40. 158. Ventura F, Ohlen J, Koinberg I. An integrative review of supportive e-health programs in cancer care. European journal of oncology nursing : the official journal of European Oncology Nursing Society. 2013;17(4):498-507. 159. Timmermann C. 'Just give me the best quality of life questionnaire': the Karnofsky scale and the history of quality of life measurements in cancer trials. Chronic illness. 2013;9(3):179-90. 160. Fantino B, Moore N. The self-reported Montgomery-Asberg Depression Rat- ing Scale is a useful evaluative tool in Major Depressive Disorder. BMC psychiatry. 2009;9:26. 161. David LS, William MCR, Gray JAM, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn't. Bmj. 1996;312. 162. Bandura. Health promotion by social cognitive means. Health Educ Behav. 2004;31(2):143-64. 163. Pettersson A, Bostrom KB, Gustavsson P, Ekselius L. Which instruments to support diagnosis of depression have sufficient accuracy? A systematic review. Nordic journal of psychiatry. 2015;69(7):497-508. 164. Puhan M, Frey M, Buchi S, Schunemann H. The minimal important difference of the hospital anxiety and depression scale in patients with chronic obstructive pulmonary disease. Health and Quality of Life Outcomes. 2008;6(1):46. 165. Eskelinen M, Korhonen R, Selander T, Ollonen P. Agreement between hopelessness/helplessness and Montgomery-Asberg Depression Rating Scale in healthy individuals and in patients with benign breast disease and breast cancer: a prospective case-control study in Finland. Anticancer Res. 2015;35(4):2215-21. 166. Julian LJ. Measures of anxiety: State-Trait Anxiety Inventory (STAI), Beck Anxiety Inventory (BAI), and Hospital Anxiety and Depression Scale-Anxiety (HADS-A). Arthritis care & research. 2011;63 Suppl 11(0 11):S467-S72. 167. Armento MEA, Hopko DR. The Environmental Reward Observation Scale (EROS): Development, Validity, and Reliability. Behavior Therapy. 2007;38(2):107-19. 168. Fayers PM, Bjordal K, Groenvold M, Curran D, Bottomley A. The EORTC QLQ-C30 Scoring Manual. 3rd ed. Brussels. 2001.

97 169. Cocks K, King MT, Velikova G, Martyn St-James M, Fayers PM, Brown JM. Evidence-based guidelines for determination of sample size and interpretation of the European Organisation for the Research and Treatment of Cancer Qual- ity of Life Questionnaire Core 30. J Clin Oncol. 2011;29(1):89-96. 170. Cella D, Eton DT, Lai J-S, Peterman AH, Merkel DE. Combining Anchor and Distribution-Based Methods to Derive Minimal Clinically Important Differ- ences on the Functional Assessment of Cancer Therapy (FACT) Anemia and Fatigue Scales. Journal of Pain and Symptom Management. 2002;24(6):547- 61. 171. Hooton JWL. Randomization tests: statistics for experimenters. Computer Methods and Programs in Biomedicine. 1991;35(1):43-51. 172. Graneheim UH, Lundman B. Qualitative content analysis in nursing research: concepts, procedures and measures to achieve trustworthiness. Nurse Educa- tion Today. 2004;24(2):105-12. 173. Trials of War Criminals before the Nuremberg Military Tribunals under Con- trol Council Law (1949). [20-02-26]. Available from: https://history.nih.gov/research/downloads/nuremberg.pdf. 174. Association WM. World Medical Association Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects. JAMA. 2013;310(20):2191-4. 175. Mohandas H, Jaganathan SK, Mani MP, Ayyar M, Rohini Thevi GV. Cancer- related fatigue treatment: An overview. Journal of cancer research and thera- peutics. 2017;13(6):916-29. 176. Israilov S, Cho HJ. How Co-Creation Helped Address Hierarchy, Over- whelmed Patients, and Conflicts of Interest in Health Care Quality and Safety. AMA journal of ethics. 2017;19(11):1139-45. 177. Fleming T, Merry S, Stasiak K, Hopkins S, Patolo T, Ruru S, et al. The Im- portance of User Segmentation for Designing Digital Therapy for Adolescent Mental Health: Findings From Scoping Processes. JMIR mental health. 2019;6(5):e12656. 178. Alsem MW, van Meeteren KM, Verhoef M, Schmitz MJWM, Jongmans MJ, Meily-Visser JMA, et al. Co-creation of a digital tool for the empowerment of parents of children with physical disabilities. Research Involvement and En- gagement. 2017;3(1):26. 179. van den Berg SW, Gielissen MFM, Custers JAE, van der Graaf WTA, Ot- tevanger PB, Prins JB. BREATH: Web-Based Self-Management for Psycho- logical Adjustment After Primary Breast Cancer—Results of a Multicenter Randomized Controlled Trial. Journal of Clinical Oncology. 2015;33(25):2763-71. 180. Willems RA, Bolman CA, Mesters I, Kanera IM, Beaulen AA, Lechner L. Short-term effectiveness of a web-based tailored intervention for cancer survivors on quality of life, anxiety, depression, and fatigue: randomized controlled trial. Psychooncology. 2017;26(2):222-30. 181. Willems RA, Mesters I, Lechner L, Kanera IM, Bolman CAW. Long-term effectiveness and moderators of a web-based tailored intervention for cancer survivors on social and emotional functioning, depression, and fatigue: randomized controlled trial. Journal of cancer survivorship : research and practice. 2017;11(6):691-703.

98 182. Rhondali W, Freyer G, Adam V, Filbet M, Derzelle M, Abgrall-Barbry G, et al. Agreement for depression diagnosis between DSM-IV-TR criteria, three validated scales, oncologist assessment, and psychiatric clinical interview in elderly patients with advanced ovarian cancer. Clinical interventions in aging. 2015;10:1155-62. 183. Hopko DR, Armento ME, Robertson SM, Ryba MM, Carvalho JP, Colman LK, et al. Brief behavioral activation and problem-solving therapy for depressed breast cancer patients: randomized trial. J Consult Clin Psychol. 2011;79(6):834-49. 184. Lupton D. The digitally engaged patient: Self-monitoring and self-care in the digital health era. Social Theory & Health. 2013;11(3):256-70. 185. Rood JA, Van Zuuren FJ, Stam F, van der Ploeg T, Huijgens PC, Verdonck- de Leeuw IM. Cognitive coping style (monitoring and blunting) and the need for information, information satisfaction and shared decision making among patients with haematological malignancies. Psychooncology. 2015;24(5):564- 71. 186. Simard S, Thewes B, Humphris G, Dixon M, Hayden C, Mireskandari S, et al. Fear of cancer recurrence in adult cancer survivors: a systematic review of quantitative studies. Journal of cancer survivorship : research and practice. 2013;7(3):300-22. 187. Thewes B, Butow P, Bell ML, Beith J, Stuart-Harris R, Grossi M, et al. Fear of cancer recurrence in young women with a history of early-stage breast cancer: a cross-sectional study of prevalence and association with health behaviours. Supportive care in cancer : official journal of the Multinational Associ- ation of Supportive Care in Cancer. 2012;20(11):2651-9. 188. Saphner T, Tormey DC, Gray R. Annual hazard rates of recurrence for breast cancer after primary therapy. J Clin Oncol. 1996;14(10):2738-46. 189. van de Wal M, Thewes B, Gielissen M, Speckens A, Prins J. Efficacy of Blended Cognitive Behavior Therapy for High Fear of Recurrence in Breast, Prostate, and Colorectal Cancer Survivors: The SWORD Study, a Randomized Controlled Trial. J Clin Oncol. 2017;35(19):2173-83. 190. Song L, Rini C, Deal AM, Nielsen ME, Chang H, Kinneer P, et al. Improving couples' quality of life through a Web-based prostate cancer education intervention. Oncology nursing forum. 2015;42(2):183-92. 191. McCaughan E, McKenna H. Never-ending making sense: towards a substan- tive theory of the information-seeking behaviour of newly diagnosed cancer patients. Journal of Clinical Nursing. 2007;16(11):2096-104. 192. van Mierlo T. The 1% rule in four digital health social networks: an observa- tional study. J Med Internet Res. 2014;16(2):e33. 193. Sun N, Rau PP-L, Ma L. Understanding lurkers in online communities: A literature review. Comput Hum Behav. 2014;38:110-7. 194. Signorelli GR, Lehocki F, Mora Fernandez M, O'Neill G, O'Connor D, Bren- nan L, et al. A Research Roadmap: Connected Health as an Enabler of Cancer Patient Support. J Med Internet Res. 2019;21(10):e14360. 195. Alberts NM, Hadjistavropoulos HD, Titov N, Dear BF. Patient and provider perceptions of Internet-delivered cognitive behavior therapy for recent cancer survivors. Supportive care in cancer : official journal of the Multinational As- sociation of Supportive Care in Cancer. 2018;26(2):597-603. 196. Dirkse D, Hadjistavropoulos HD, Alberts NA, Karin E, Schneider LH, Titov N, et al. Making Internet-delivered cognitive behaviour therapy scalable for cancer survivors: a randomized non-inferiority trial of self-guided and techni- cian-guided therapy. Journal of Cancer Survivorship. 2019.

99 197. DuBenske LL, Gustafson DH, Namkoong K, Hawkins RP, Atwood AK, Brown RL, et al. CHESS improves cancer caregivers' burden and mood: results of an eHealth RCT. Health Psychol. 2014;33(10):1261-72. 198. Köhle N, Drossaert CHC, Schreurs KMG, Hagedoorn M, Verdonck-de Leeuw IM, Bohlmeijer ET. A web-based self-help intervention for partners of cancer patients based on Acceptance and Commitment Therapy: a protocol of a randomized controlled trial. BMC Public Health. 2015;15(1):303. 199. Richards DA, Borglin G. ‘Shitty nursing’ – The new normal? International Journal of Nursing Studies. 2019;91:148-52. 200. Wise M, Han JY, Shaw B, McTavish F, Gustafson DH. Effects of using online narrative and didactic information on healthcare participation for breast cancer patients. Patient Educ Couns. 2008;70(3):348-56. 201. Danaher BG, Seeley JR. Methodological Issues in Research on Web-Based Behavioral Interventions. Annals of Behavioral Medicine. 2009;38(1):28-39. 202. Langius-Eklöf A, Christiansen M, Lindström V, Blomberg K, Hälleberg Ny- man M, Wengström Y, et al. Adherence to Report and Patient Perception of an Interactive App for Managing Symptoms During Radiotherapy for Prostate Cancer: Descriptive Study of Logged and Interview Data. JMIR cancer. 2017;3(2):e18-e. 203. Schulz KF, Altman DG, Moher D. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. Bmj. 2010;340:c332. 204. Urech C, Grossert A, Alder J, Scherer S, Handschin B, Kasenda B, et al. Web- Based Stress Management for Newly Diagnosed Patients With Cancer (STREAM): A Randomized, Wait-List Controlled Intervention Study. J Clin Oncol. 2018;36(8):780-8. 205. Ekholm A. Bortom IT : om hälsa i en digital tid. Stockholm: Institutet för fram- tidsstudier; 2016. 206. Friberg AS, Rask Moustsen I, Benzon Larsen S, Hartung T, Wreford Andersen E, Halgren Olsen M, et al. Educational level and the risk of depression after prostate cancer. Acta oncologica. 2019;58(5):722-9. 207. Dalton SO, Olsen MH, Johansen C, Olsen JH, Andersen KK. Socioeconomic inequality in cancer survival – changes over time. A population-based study, Denmark, 1987–2013. Acta oncologica. 2019;58(5):737-44. 208. Burrell AMG, Allan JL, Williams DM, Johnston M. What do self-efficacy items measure? Examining the discriminant content validity of self-efficacy items. British Journal of Health Psychology. 2018;23(3):597-611. 209. Merluzzi TV, Martinez Sanchez MA. Assessment of self-efficacy and coping with cancer: Development and validation of the Cancer Behavior Inventory. Health Psychology. 1997;16(2):163-70. 210. Ryan C, Bergin M, Wells JSG. Theoretical Perspectives of Adherence to Web- Based Interventions: a Scoping Review. International Journal of Behavioral Medicine. 2018;25(1):17-29. 211. Christensen H, Mackinnon A. The law of attrition revisited. J Med Internet Res. 2006;8(3):e20-e1. 212. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ : British Medical Journal. 2014;348:g1687. 213. Malterud K, Siersma VD, Guassora AD. Sample Size in Qualitative Interview Studies: Guided by Information Power. Qual Health Res. 2015. 214. Kvale S, Brinkmann S. InterViews : learning the craft of qualitative research interviewing. Los Angeles: Sage Publications; 2009.

100 215. Polit DF, Beck CT. Nursing research : generating and assessing evidence for nursing practice. 9 ed: Philadelphia: Wolters Kluwer Health/Lippincott Wil- liams & Wilkins; 2012. p. 582-6. 216. Stemler S. An Overview of Content Analysis. Practical Assessment, Research & Evaluation. 2001;7(17):1-6. 217. Renz SM, Carrington JM, Badger TA. Two Strategies for Qualitative Content Analysis: An Intramethod Approach to Triangulation. Qualitative Health Re- search. 2018;28(5):824-31.

101 Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1667 Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-408909 2020