- Email: [email protected]
Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers?
YEBEH-05472; No of Pages 7 Epilepsy & Behavior xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Maria Grazia Celani a,b,⁎, Teresa Anna Cantisani b,c, Angelo Bignamini d, Sara Macone b,e, Rossella Papetti c, Luisa Baiocco c a
Department of Neurophysiopathology, Azienda Ospedaliera di Perugia, Ospedale S. Maria della Misericordia, S. Andrea delle Fratte, 06156 Perugia, Italy Cochrane Neurosciences Field, Direzione Salute Welfare, Region of Umbria, Perugia, Italy Perugia Hospital, Neurophysiopathology Unit, Azienda Ospedaliera di Perugia, Ospedale S. Maria della Misericordia, S. Andrea delle Fratte, 06156 Perugia, Italy d University of Milan, Department of Pharmacology, Milan, Italy e University of Kansas Medical Center, Department of Dietetics and Nutrition, USA b c
a r t i c l e
i n f o
Article history: Received 15 June 2017 Revised 19 August 2017 Accepted 19 August 2017 Available online xxxx Keywords: Epilepsy Evidence-based neurology Quality of life Psychology Outcome measures
a b s t r a c t Objective: The objective of this study was to collect perspectives, ideas, and values of people affected by epilepsy and their carer to include them in new research in this area. Design: This is a semiquantitative study analyzing needs, emotions, and medical issues emerging from focus groups with patients and carers divided in three severity levels of disease. Participants: The participants were 25 patients and 36 carers attending outpatients' service of tertiary epilepsy center in Umbria, Italy. Results: Assistance was the need expressed by more than 40% of the participants followed by experience-sharing, knowledge, control, clarity, and security. The only significant effect in logistic regression analysis after adjusting for severity was the patients' expressed need for “experience-sharing” more than their carers (OR 7.29, 95%CI: 1.76–30.18). Hope was the emotion expressed by more than 50% of the participants, followed by anger, fear, and resignation. After adjusting for severity, carers were more likely to express anger, in comparison with patients (OR 17.23, 95%CI: 3.55–83.74; P b 0.001). The patients were 6.88 times more likely (95%CI 1.84–25.75; P = 0.004) to express “resignation” than their carers. The most frequent medical issues were related to: “medications”, “frequency of crises”, “work impact”, “quality of life”, “psychomedical integration of care”, and “development of new drugs”. After adjusting for severity in a logistic regression analysis, patients were more likely to express concerns for the frequency of crises than carers (OR 3.57, 95%CI: 1.16–11.04; P = 0.027). Conclusions: Patients' and carers' priorities, based on intense personal insight, represent a starting point to work for shared outcome measures in clinical trials and shared agenda in research, including research of strong evidence in complex intervention as service models for people with epilepsy. © 2017 Elsevier Inc. All rights reserved.
1. Introduction A sensible research agenda must consider the needs of patients and caregivers, alongside those of clinicians, researchers, decision makers, and even the pharmaceutical industry so that each stakeholder is aware of the others' needs [1]. Academic institutions have generally left the choice of topics and methodologies to their professional and scientific committees, often ignoring nonprofessionals when making these decisions. Patients' preferences should be included in the choice ⁎ Corresponding author at: Department of Neurophysiopathology, Azienda Ospedaliera di Perugia, Ospedale S. Maria della Misericordia, S. Andrea delle Fratte, 06156 Perugia, Italy. E-mail addresses: [email protected], neurofi[email protected] (M.G. Celani).
of topics and in the design of clinical trials, especially regarding the selection of outcome measures, the most relevant and important part of research for the end users of treatments and care [2]. Seizure frequency, in epilepsy research, is often the primary outcome in clinical intervention trials; however, it may not be the best predictor of patients' daily life skills, and it cannot be assumed that good psychosocial outcomes will be associated with the attainment of a seizure-free state. It may be necessary to go beyond seizure counts in the assessment of the disease and beyond the simple appearances of symptoms. Similarly, the obsessive research for a new drug may not be the exclusive aspiration of a patient. It could be useful to look for further and innumerable aspects that undermine the life of the ill individuals and their carer [3]. The aim of this work was to elicit perspectives, ideas, and values from people with epilepsy and their carers in order to offer clinicians
http://dx.doi.org/10.1016/j.yebeh.2017.08.028 1525-5050/© 2017 Elsevier Inc. All rights reserved.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
2
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
and researchers new indications, suggestions, and hints able to reshape the future of treatment and care in a joint effort. 2. Materials and methods 2.1. Inclusion criteria Adult patients affected by epilepsy, attending the outpatient service of the Neurophysiopathology Department, Santa Maria Misericordia Hospital, Perugia, Italy, and carers were invited to participate in this study. Patients had to be over 18 years of age, able to give informed consent, and have a clinical diagnosis of epilepsy. Caregivers, including all the people involved in the care of the patient: family members, friends, significant others, and personal home caretakers, had to be over 18 years of age and able to give informed consent. 2.2. Exclusion criteria Participants who were unable to provide informed consent or were affected by learning difficulties and medium to severe mental challenges were excluded.
2.3. Methodology Patients and carers were recruited by phone using the patients' lists from the outpatient epilepsy clinic. During recruitment, 15 phone call invitations were made by a researcher with the aim of having 6 to 8 participants in each group Fig. 1. During each phone call, the researcher informed the potential participant of the aim, methodology, and practical aspects of the study. During the conversation, an evaluation of the patient was made. Participants who agreed to take part in the study were divided into different groups based on the severity of his/her disease and were classified, using the different grades of the Modified Rankin Scale 0–2, 3, 4–5, into three different groups: mild, moderate, and severe. This scale is a simple, easy-to-administer tool, and can be given over the phone by both experts and nonmedical study personnel. It is used to measure functional outcome in a broad population of patients with neurological diseases [4–6]. Patients in the mild group are those who are self-sufficient, with a good pharmacological control of their seizures, and no loss of personal or social abilities. Moderate patients often require the assistance of a caretaker in order to accomplish some of their personal and social activities. Severe patients depend almost completely on a caregiver for every need. Once each participant with epilepsy was
Fig. 1. Flow chart patients' and carers' responses.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
categorized into the appropriate group, his/her caregiver (or the caregiver of another patient with the same characteristics) was included in the equivalent group. Two groups of patients and two groups of caregivers were formed for each level of severity of disease, except for the severe patients (one group only) for a total of 11 focus groups (Table 1). In order to obtain and measure patient-centered objectives, we used focus groups, a qualitative type of methodology used in health research [7]. There were 4 to 10 people in each focus group; all groups discussed the same topic, answering predetermined and semistructured questions (see Appendix 1 for focus group schedules), and each session was recorded. Two psychologists (RP and LB), trained in engaging in interactive group settings, conducted the recruitment, obtained verbal and written consensus to participate, and were the moderators of the focus groups. They guided the opening of the discussions and regulated the session so that all participants had the opportunity to actively participate. One moderator asked the questions and regulated response times, while another managed the recording, the moderators also helped to stay on topic, balance group participation to help make all group members feel comfortable. Focus groups averaged 90 min. At the end of each session, the moderator gave a short summary which included all the points of view that had emerged during discussion to make sure that each participant felt their contribution to the discussion that had been considered. The recorded group discussions were transcribed into text files. The transcripts were independently read by the moderators in order to blindly identify the key semantic meanings expressing the patients' perceived needs, as well as the emotions regarding the disease and its consequences. Two neurologists (MGC and TAC) also examined the files to highlight health and medical-related matters. After this first step, the findings were discussed to compare results and identify equivalent key themes from participants' reports. 2.4. Data analysis Each transcript was analyzed with “Concordance” software (version 3.3, University of Dundee, UK Watt R.J.C) using disease severity as classifier (mild, moderate, and severe), the subject's severity status (pt/c), the language used (formal language or dialect), and, for carers, the proximity of their relationship to the patient. Word frequency was computed after removing the usual stop words (articles, prepositions, etc.). The number of words by subject did include the stop words. For this analysis, only the following classifiers were used: severity and status. Needs, emotions, and medical-related matters were also classified for future semantic analyses; for this analysis, however, they were transformed into headwords and counted with software facilities. The results stratified by classifiers were then exported as “.xml files” and loaded into R and SPSS for statistical analysis. The descriptive analysis used mean ± (SD) Standard Deviation and/ or 95% Confidence Interval (95%CI) for normally distributed variables, median with 95%CI estimated by bootstrapping for nonnormally distributed variables, proportions with 95%CI as appropriate, for nominal variables. Comparisons made use of the t-test and analysis of variance (ANOVA) for normally distributed variables; Mann–Whitney and Kruskal–Wallis
3
tests for variables significantly deviating from normality, with the chi square test, the chi square test for trend, the exact Fisher test, and/or the logistic regression analysis for nominal variables. The ANOVA of variables significantly deviating from the assumption of normality was performed by bootstrapping using the unrestricted permutation of observations [8], rather than with the rank ANOVA. 3. Results We analyzed the responses of 25 patients and 36 carers, of whom 31 first-degree relatives (mother/father; sister/brother) and 5 spouses (Table 1). Women represented 50% of carers and 56% of patients (P = 0.644); gender distribution was not modified by severity (P = 0.537). As expected, age was significantly higher among carers (median: 65; range: 49–76) than patients (median: 40; range :19–76; P b 0.001, Mann–Whitney test) and this difference was homogeneous throughout severity levels (P = 0.469; Kruskal–Wallis test). The first item examined was whether the participants spoke about themselves. To this aim, we examined the frequency of the words related to “I” (I, me, my, mine, etc. [Table 2]). Being “a patient” determined a significantly greater tendency to speak about themselves, in comparison with being “a carer”. The severity of disease did not influence how often patients referred to themselves during the focus groups, and there was no evidence of interaction between severity and status. The model that excluded the interaction term showed a significant effect of the status (P b 0.001) and not of its severity (P = 0.323). 3.1. Analysis of classifiers 3.1.1. Propensity to express The predisposition of participants to express needs (status = 0.305; severity = 0.083), emotions (status = 0.138; severity = 0.089), or health-related issues and medical matters (status = 0.998; severity = 0.820) did not differ between pt/c or by severity of disease (data not shown). Monitoring instead the density of expression (average number of words uttered per need, emotion or medical issue expressed), patients' speech was significantly denser in needs and medical issues regardless of severity, while the density in emotions increased with severity, regardless of status (Table 3). 3.1.2. Expressed needs Overall, 12 different needs were expressed 169 times by the 61 participants (pt/c). More than 40% of the participants asked for “assistance” i.e., home health care, psychological support, a more intimate and committed patient–doctor relationship, better use of skilled social and health workers, and specific resources for patients with epilepsy. In terms of total number of expressions of needs, the most relevant ones were assistance, experience-sharing, knowledge, control, clarity, and security. If we assume that the “mean number of expressions per subject expressing” represents the “relevance for the individual”, their most relevant expressed needs were assistance, experience-sharing, knowledge, and outpouring. The subject expressing the word “assistance” repeated it more than 2.3 times (Table 4).
Table 1 Participants in focus group divided by status and disease severity. Disease severity by Modified Rankin Scale
Mild (0–2)
Moderate (3)
Severe (4–5)
Number of patients in each focus group
FOCUS n. 1:4 FOCUS n. 7:7 11 FOCUS n. 2:4 FOCUS n. 8:6 10
FOCUS n. 3:4 FOCUS n. 9:6 10 FOCUS n. 4:8 FOCUS n. 10:7 15
FOCUS n. 5:4 4 FOCUS n. 6:5 FOCUS n. 11:6 11
TOT Number of carer in each focus group TOT Numbers in bold are participants.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
4
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
Table 2 Mean percent proportion ± SD (N) of terms related to “I” on the total number of significant words pronounced. Mild
Moderate
Severe
Table 4 Distribution of expressed needs. Expressed needs
n. of subjects expressing (% of total subjects)
n. of expressions (% of total expressions)
Mean ± SD n. of expressions per subject expressing
Assistance Experience-sharing Knowledge Clarity Control Planning Security Comprehension Predictability Outpouring Doing Respect Total
26 (42.6%) 14 (23.0%) 12 (19.7%) 9 (14.8%) 6 (9.8%) 6 (9.8%) 5 (8.2%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 2 (3.3%) 1 (1.6%) 61
60 (35.5%) 32 (18.9%) 25 (14.8%) 10 (5.9%) 11 (6.5%) 7 (4.1%) 9 (5.3%) 3 (1.8%) 3 (1.8%) 6 (3.6%) 2 (1.2%) 1 (0.6%) 169
2.31 ± 1.57 2.29 ± 2.20 2.08 ± 1.08 1.11 ± 0.33 1.83 ± 0.75 1.17 ± 0.41 1.80 ± 0.84 1.00 ± 0.00 1.00 ± 0.00 2.00 ± 1.00 1.00 ± 0.00 1.00 ± − 1.88
Total
Carers 2.49 ± 1.32 (10) 2.64 ± 0.91 (15) 2.76 ± 1.92 (11) 2.64 ± 1.36 (36) Patients 5.62 ± 2.26 (11) 4.16 ± 1.13 (10) 5.80 ± 0.48 (4) 5.06 ± 1.79 (25) Total 4.13 ± 2.43 (21) 3.25 ± 1.24 (25) 3.57 ± 2.15 (15)
More patients than carers expressed the need for “experiencesharing” (40.0%, 95%CI: 21.1%–61.3% vs. 11.1% 95%CI: 3.1%–26.1%; Fisher test P = 0.013) and “control” (20.0%, 95%CI: 6.8%–40.7% vs. 2.8%, 95%CI: 0.0%–14.5%; P = 0.038). No other significant difference was seen for the other needs expressed. To estimate the effect of status (pt/c) and severity of disease on the distribution of the most frequently expressed needs, logistic regression analysis was applied. The only significant effect was seen by status (pt/c) on experience-sharing (P = 0.006). After adjusting for severity, patients expressed the need for experience-sharing more than their carers (OR 7.29; 95%CI: 1.76%–30.18%). If we consider patients and carers together, stratified by disease severity, the need for planning (P = 0.010; chi square for trend) and for assistance (P = 0.227, NS) was directly related to disease severity while the need for control was inversely correlated (P = 0.054, NS) (Fig. 2). In other words, the need for control was more relevant for people with a mild disease while, as expected, the need for planning and assistance was important in people with more disabilities. 3.1.3. Expressed emotions Overall, 20 different emotions were expressed 341 times by the 61 participants (pt/c). More than 50% of the participants expressed “hope”, i.e., faith, having a normal life, more effective medications, and decrease in seizure frequency. In terms of total number of expressed emotions, the most relevant were hope, anger, fear, and resignation. In terms of relevance for each participant, anger, resignation, fear, anxiety/apprehension, acceptance, hope, and shame ranged from 2.36 to 1.56 (Table 5). The frequency of participants expressing “resignation” at least once was greater among patients than carers (respectively 48.0%, 95%CI: 27.8%–68.7% vs. 13.9%, 95%CI: 4.7%–29.5%; Fisher test P = 0.008). The frequency of “disappointment” followed a similar trend (12.0%, 95%CI: 2.5%–31.2% vs. 0.0%, 95%CI: 0.0%–9.7%; P = 0.064). On the contrary, the frequency of participants expressing “anger” was much greater among carers than patients (61.1%, 95%CI: 43.5%–76.9% vs. 12.0%,
95%CI: 2.5%–31.2%; P b 0.001). No other significant difference was seen for the other emotions expressed. The distribution of the most frequently expressed emotions was tested by logistic regression analysis to estimate the effect of status and severity. The only significant effects were by status on “anger” and on “resignation”. After adjusting for severity, carers were more likely to express anger, in comparison with patients (OR 17.23, 95%CI 3.55–83.74; P b 0.001). Patients were 6.88 times more likely (95%CI 1.84–25.75; P = 0.004) to express “resignation” than their carers. Considering the frequency of repetition per subject, there was no particular emotion that was repeated more frequently. When stratified by disease severity, the proportion of participants (pt/c) expressing shame was borderline associated with increased severity (P = 0.051), while in terms of proportion of responses made by participants, the relevance of “hope” (P = 0.002 chi square for trend) and “shame” (P = 0.002) highly increased, while “acceptance” (P = 0.003) decreased with severity of disease (Fig. 3).
3.1.4. Health-related issues and medical matters — physicians' perspectives Overall, 25 different medical topics were expressed 358 times by the 61 participants. The most frequent ones were related to “medications”, “frequency of crises”, “work impact”, “quality of life”, “psychomedical integration of care”, and “development of new drugs”. Among the total number of expressed topics, the most relevant ones were “work impact”, “medications”, “new drugs”, and “drug adverse reactions” (Table 6).
Table 3 Median number of words (±95%CI) pronounced per need, emotion or medical issues expresseda. Item Needs
Emotions
Medical issues
a b c
Carers Patients Total Pb Pc Carers Patients Total Pb Pc Carers Patients Total Pb Pc
456 [206–1261] 565 [440–759] 213 [144–321] 297 [231–393] 240 [185–451] 440 [310–565] 0.09195 Status = 2e−04; severity = 0.786; interaction = 0.832 146 [93–207] 232 [217–257] 102 [71–120] 231 [126–306] 107 [89–145] 232 [197–270] 0.00071 Status = 0.1486; severity = 0.010; interaction = 0.679 78 [49–176] 78 [64–144] 40 [27–67] 51 [37–85] 52 [40–78] 68 [56–90] 0.27204 Status = 0.009; severity = 0.783; interaction = 0.134
413 [208–694] 116 [88–148] 250 [133–416]
456 [378–630] 213 [148–310]
0.00025
139 [76–416] 130 [99–177] 139 [99–189]
196 [152–228] 120 [99–155]
0.08754
63 [33–187] 40 [28–216] 47 [33–123]
74 [63–119] 46 [38–56]
0.00113
Mann–Whitney test for status. Kruskal–Wallis test for severity. ANOVA (bootstrapping, 5000 replications).
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
5
Fig. 2. Proportion of subject expressing a need for planning assistance and control by disease severity.
Fifty-six percent of the patients (95%CI: 34.9%–75.6%) expressed concern for drugs at least once compared with 22.2% of the carers (95%CI 10.1%–2801% Fisher exact test P = 0.014); “concern to adjustment of new lifestyle” 28.0% (95%CI 12.1%–49.4%) vs. 5.6% (95%CI 0.7%–18.7%) P = 0.024, and “fear of seizure recurrence” 52.0%(95%CI 31.3%–72.2%) vs. 22.2% (95%CI 10.1%–39.2%) P = 0.025. On the contrary, carers more frequently than patients expressed greater “concern for work” (44.4%; 95%CI 27.9%–61.9% vs. 20.0%; 95%CI 6.8%–40.7%; P = 0.059 borderline to significance). No other significant differences were seen. The distribution of the issues expressed more frequently (medications, frequency of crises, work, and quality of life) was also tested by using logistic regression analysis, in order to estimate the effect of status (pt/c) and severity adjusted for each category. The only statistically significant effect was for patients who were more likely to express concerns for “fear of seizure recurrence” than carers (OR 3.57; 95%CI 1.16–11.04; P = 0.027). No other significant differences were seen.
Table 5 Distribution of expressed emotions. Emotion
n. of participants expressing (% of total participants)
n. of expressions (% of total expressions)
Mean ± SD n. of expressions per subject expressing
Hope Anger Fear Sorrow Resignation Acceptance Anxiety/Apprehension Shame Bewilderment Guilt Disappointment Detachment Remorse Happiness/Serenity Reject Curiosity Gratitude Exasperation Surprise Sarcasm Total
33 (54.1%) 25 (41.0%) 30 (49.2%) 22 (36.1%) 17 (27.9%) 17 (27.9%) 14 (23.0%) 9 (14.8%) 7 (11.5%) 4 (6.6%) 3 (4.9%) 2 (3.3%) 1 (1.6%) 2 (3.3%) 2 (3.3%) 1 (1.6%) 1 (1.6%) 1 (1.6%) 1 (1.6%) 1 (1.6%) 61
60 (17.6%) 59 (17.3%) 59 (17.3%) 27 (7.9%) 35 (10.3%) 32 (9.4%) 27 (7.9%) 14 (4.1%) 8 (2.3%) 4 (1.2%) 3 (0.9%) 2 (0.6%) 2 (0.6%) 2 (0.6%) 2 (0.6%) 1 (0.3%) 1 (0.3%) 1 (0.3%) 1 (0.3%) 1 (0.3%) 341
1.82 ± 1.69 2.36 ± 2.00 1.97 ± 1.35 1.23 ± 0.43 2.06 ± 1.60 1.88 ± 1.17 1.93 ± 1.94 1.56 ± 0.53 1.14 ± 0.38 1.00 ± 0.00 1.00 ± 0.00 1.00 ± 0.00 2.00 ± − 1.00 ± 0.00 1.00 ± 0.00 1.00 ± − 1.00 ± − 1.00 ± − 1.00 ± − 1.00 ± − 1.77
4. Discussion The real novelty of our study is that by using the text analysis technique we turned qualitative data into semiquantitative ones. Two professional blinded experts indexed the content of focus groups, separately coded important terms, and categorized open-ended common responses. Concordance between the two coders was 90%. By doing so, we reduced the gap between what people actually said in focus groups and what the moderator interpreted in order to minimize interpretative biases. We chose, as the primary tool of our research, focus groups' discussions, in order to freely elicit how people feel, function, share, and compare their different points of views instead of the more predetermined self-rated questionnaires. We grouped participants by severity of disease and by their status, i.e., being a patient or a carer, with the aim to understand the real differences in perspectives between patients and family members and the way these perspectives change depending on the severity of the disease. We confirmed our hypothesis that there is a large difference in perspective between patients and carers. The amount of words related to “I” (I, me, my, mine) that we coded as “self-expressed” suggests that patients speak of and for themselves, whereas carers speak on behalf of the patients they care for. We speculate that this peculiar result depends on a cultural and religious attitude of moral awareness (strong sense of duty) and on the ethical responsibility in looking after people [9]; for this reason, we cannot exclude that this feeling is particularly strong in Italy where there is a historical imprinting of Catholicism and might reflect the small and homogeneous sample size. Furthermore, patients tend to go straight to the issue with few words to express a need or medical aspect, while carers need significantly more. Another aspect of our study is about what we refer to as “needs”. The most relevant “needs” that emerged from all participants were “assistance”, “experience-sharing”, “knowledge”, “clarity”, “control”, and “planning”. “Assistance”, considered as the access to tailored and more dedicated health services, is the most specific need expressed by all participants, while “experience-sharing” and “knowledge” are unique needs expressed particularly by patients. People affected with the disease wish to have the opportunity to compare with other patients, support each other, and have a deeper knowledge of the medical aspects of epilepsy in order to have better control of the disease and medications, and consequently, reduce fear. In this study, the need defined as “planning” is considered the necessity of having easier access to employment possibilities and social life, from which patients with epilepsy are often excluded. “Planning” is strongly correlated to severity of disease, while “Control”, intended as the possibility to manage the different aspects of the disease, like related risk of seizures, medication
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
6
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
Fig. 3. Statements expressing hope, shame, and acceptance estimated as a proportion of the total statements expressing an emotion stratified by disease severity.
side effects, the unpredictability of the seizure event, and controlling families and carers, is inversely proportional to severity. We also explored participants' “emotions”, in relation to the disease. In this regard, patients expressed resignation and failure (i.e., “I have accepted it”, “my life had to slow down”, “I had to give up professional soccer”) while carers expressed mostly anger (“the whole family and our routine was effected by the diagnosis, and the relationship with the ill person is often hard because his/hers life is ruined, and so is ours”). Surprisingly, alongside the negative emotions, participants expressed “hope”, wishing for a “normal” life, more effective medications, and decrease of seizure frequency. We found a marginal statistically significant emotion regarding “shame” among severely affected people, and we were consequently able to speculate that there is still the perception among patients of being rejected and most of all stigmatized by society [10]. Stigma still seems to be one of the most hurtful and persistent issues. Kale more than 20 years ago stated that “the history of epilepsy can be summarized as 4000 years of ignorance, superstition, and stigma, followed by
100 years of knowledge, superstition, and stigma”. While conducting our study, we came across the fact that very little has changed regarding this disease, despite the numerous initiatives, including the famous joint project called “Out of shadow”, started in 1997 [11]. Stigmatization leads to discrimination, and people affected by epilepsy have been more often than not targeted throughout history and cultures with prejudicial behaviors. When the two physicians interpreted the focus group texts, they deduced that approximately 20% of the total expressions by participants were about epileptic seizures (number, frequency, length) while 24.6% were about medication (drugs in general, new medications, adverse events). A concerned spouse and caregiver stated that because of the psychoactive effects of the medication, he no longer had any idea of who his wife actually was. A straightforward consequence, given the effective cognitive behavioral effects of medications and the related concerns of consumers, is the necessity to conduct prolonged follow-ups throughout the years in order to observe psychological variations of personality. 4.1. Limitations
Table 6 Distribution of expressed medical issues. Medical issue
No. subjects expressing (% of total subjects)
No. expression (% of total expressions)
Drugs Frequency of crises in general Work Quality of life Psych medical cure integration New drugs Medical assistance Small crises intensity Crises intensity Adverse reactions Hobbies Proper information Driving license Frequency of severe crises Education to new lifestyle Frequency of small crises Social interaction Genitoriality Drug dependence/need Psychiatric comorbidity Drug efficacy without ADRs Surgery Individualized project Intercritic time Drug cost Total
22 (36.1%) 21 (34.4%) 21 (34.4%) 17 (27.9%) 16 (26.2%) 16 (26.2%) 15 (24.6%) 14 (23.0%) 13 (21.3%) 12 (19.7%) 12 (19.7%) 12 (19.7%) 12 (19.7%) 10 (16.4%) 9 (14.8%) 9 (14.8%) 9 (14.8%) 5 (8.2%) 5 (8.2%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 1 (1.6%) 61
27 (7.5%) 24 (6.7%) 31 (8.7%) 24 (6.7%) 23 (6.4%) 25 (7.0%) 22 (6.1%) 16 (4.5%) 17 (4.7%) 25 (7.0%) 15 (4.2%) 16 (4.5%) 15 (4.2%) 10 (2.8%) 10 (2.8%) 9 (2.5%) 14 (3.9%) 6 (1.7%) 7 (2.0%) 4 (1.1%) 4 (1.1%) 4 (1.1%) 6 (1.7%) 3 (0.8%) 1 (0.3%) 358
This study was carried out among a population with epilepsy attending the outpatient clinic of a tertiary epilepsy center in a single Italian region. Therefore, it may reflect the cultural and the social aspects of this geographical area, along with the characteristics of the local healthcare system and the environment where the participants live. We have taken into account that some aspects of our research may not apply to different contexts, even with a similar reality. Results could differ even more when compared to care and disease management in developing countries where the “organization of care” must be adapted to the existing healthcare system [12]. The major limitations of this study are the small sample size and the fact that participation to the focus groups was fully voluntary. People attending the focus groups were about one-third of those contacted. Taking into account that participation required time and effort, it is likely that the participants were those most motivated, consequently, these results might not be extendable to the overall population. 4.2. Strengths This is one of the few studies that analyze the precious information that comes from listening to people who carry the burden of disease. The analysis was carried out using a semiqualitative method to guarantee that the expressions used cannot be manipulated. This perspective offers the possibility of bridging the gap that exists between people who receive funding and those who make the decision regarding future
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
research and what patients, families, and the public would like to see explored. This would help us to move toward patient-centered medicine and tailored outcome measures, implemented in a patientcentered health organization. 5. Conclusions The analysis of the focus groups shows that reduction of seizure frequency, the oldest and the most common outcome measure in epilepsy clinical trials, is still the most important for people affected by the disease. On the other hand, drug side effects, which are at best considered a secondary outcome from the main study objectives in clinical trials, become extremely important to daily living both for patients and carers. It is for this reason that we feel a primary outcome in epilepsy research that should include a sensible balance between benefits and risks at short and long term. A second aspect can be summarized with the comprehensive care concept, derived from expressed patients' and carers' needs. Patients have the desire to become empowered and knowledgeable about the correct health behaviors that need to be learned in order to better control their seizures, to better adapt themselves to the constraints of their disease and promote their well-being. In addition, people affected by the disease expect that health services will include highly specialized epilepsy care in liaison with multidisciplinary teams available in their communities. Service models for epilepsy have been evaluated in different clinical trials and synthesized in systematic reviews published in the Cochrane Library [13]. The reviews conclude that, at present, there is some evidence of effectiveness when a specialist epilepsy nurse is at hand, and self-management education is offered. Evidence on other more complex intervention types is still unclear and should be better investigated and consequently proposed to policy makers. The patients' and carers' voices, elicited in our study, drive us to organize the future agenda in epilepsy research in this direction [14]. Conflict of interest This study had no financial support and authors have no conflict of interest. Appendix 1 Topic area for focus group interviews: 1. 2. 3. 4.
Recall the first time you were ill. What did you feel when you were told you had the disease? What problems did you have to face then and now? What do you consider the most important aspects to be improved in disease care?
Appendix 2 Procedure used to classify focus groups' results 1. Each focus group was audio recorded and transcribed into a text file. 2. Each speaker was code-identified, classified by status (patient/ carer) and severity (mild/moderate/severe). 3. All transcripts were crosschecked against the recordings independently by the two moderators (the two psychologists).
7
4. Any discrepancy between transcripts was resolved by agreement between the moderators. 5. Any reference to personal identity was removed from all text files. 6. Copies of each text file with markers for status and severity were distributed to the two psychologists and the two neurologists. 7. Each reader independently marked and assigned an emotion, need, or medical issue to the individual statements in the text files. These could be part of a sentence, a whole sentence, or span over more than one sentence. A statement included one complete semantic content. 8. The readers compared each individual statement and the semantic content applied; differences were resolved by agreement. 9. The two psychologists individually marked every statement expressing any single emotion, need, or both. The two neurologists individually marked every statement expressing a health-related issue and/or medical matter. 10. Following a brief discussion, discrepancies were resolved without the use of external support as foreseen in the protocol. 11. All classifications were revised jointly by the two psychologists for needs and emotions, and by the two neurologists for health issues and medical matters. Overlapping or very similar classifications were reexamined to determine whether they could be compounded into a single classification. Reclassification and compounding was performed by agreement between readers. 12. Final classifications were inserted as markers into the transcripts. For this study, the markers were changed into headwords and elaborated using the Concordance software, they were classified by speaker, severity, and status. The version containing all markers was archived for future semantic analyses. 13. Results from Concordance were loaded into statistical software (R and SPSS) to perform the planned analyses. References [1] Liberati A. Need to realign patient-oriented and commercial and academic research. Lancet 2011;378(9805):1777–8. [2] Crowe S, Giles C. Making patient relevant clinical research a reality. BMJ 2016;355: i6627. [3] Heath I. How medicine has exploited rationality at the expense of humanity: an essay by Iona Heath. BMJ 2016;355:i5705. [4] Candelise L, Pinardi G, Aritzu E. Interview for stroke outcome assessment. Cerebrovasc Dis 1994;4(5):341–3. [5] Patel N, Rao VA, Heilman-Espinoza ER. Simple and reliable determination of the modified Rankin scale score in neurosurgical and neurological patients: the mRS9Q. Neurosurgery 2012;71(5):971–5 [discussion 975]. [6] Savio K, Della Pietra GL, Oddone E. Reliability of the modified Rankin scale applied by telephone. Neurol Int 19 2013;5(1):e2. [7] Lehoux P, Poland B, Daudelin G. Focus group research and “the patient's view”. Soc Sci Med 2006;63:2091–104. [8] Chapman & Hall. In: BFJ Manly, editor. Randomization, bootstrap, and Monte Carlo methods in biology3rd ed. ; 2007 [London]. [9] Kleinman A. Caregiving as moral experience. Lancet 2012;380(9853):1550–1. [10] De Boer HM, Mula M, Sander JW. The global burden and stigma of epilepsy. Epilepsy Behav 2008;12(4):540–6. [11] Kale R. Bringing epilepsy out of the shadows. BMJ 1997;315:2–3. [12] Birbeck GL. Epilepsy care in developing countries: part I of II. Epilepsy Curr 2010; 10(4):75–9. [13] Bradley P, Bruce L, Nigel F. Care delivery and self-management strategies for adults with epilepsy. Cochrane database of systematic reviews. The Cochrane Collaboration; 2016 [Issue 2. Art. No: CD006244]. [14] Lynch S. Measures need to capture patients' views and experiences more effectively. BMJ 2013;346:f1553.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Maria Grazia Celani a,b,⁎, Teresa Anna Cantisani b,c, Angelo Bignamini d, Sara Macone b,e, Rossella Papetti c, Luisa Baiocco c a
Department of Neurophysiopathology, Azienda Ospedaliera di Perugia, Ospedale S. Maria della Misericordia, S. Andrea delle Fratte, 06156 Perugia, Italy Cochrane Neurosciences Field, Direzione Salute Welfare, Region of Umbria, Perugia, Italy Perugia Hospital, Neurophysiopathology Unit, Azienda Ospedaliera di Perugia, Ospedale S. Maria della Misericordia, S. Andrea delle Fratte, 06156 Perugia, Italy d University of Milan, Department of Pharmacology, Milan, Italy e University of Kansas Medical Center, Department of Dietetics and Nutrition, USA b c
a r t i c l e
i n f o
Article history: Received 15 June 2017 Revised 19 August 2017 Accepted 19 August 2017 Available online xxxx Keywords: Epilepsy Evidence-based neurology Quality of life Psychology Outcome measures
a b s t r a c t Objective: The objective of this study was to collect perspectives, ideas, and values of people affected by epilepsy and their carer to include them in new research in this area. Design: This is a semiquantitative study analyzing needs, emotions, and medical issues emerging from focus groups with patients and carers divided in three severity levels of disease. Participants: The participants were 25 patients and 36 carers attending outpatients' service of tertiary epilepsy center in Umbria, Italy. Results: Assistance was the need expressed by more than 40% of the participants followed by experience-sharing, knowledge, control, clarity, and security. The only significant effect in logistic regression analysis after adjusting for severity was the patients' expressed need for “experience-sharing” more than their carers (OR 7.29, 95%CI: 1.76–30.18). Hope was the emotion expressed by more than 50% of the participants, followed by anger, fear, and resignation. After adjusting for severity, carers were more likely to express anger, in comparison with patients (OR 17.23, 95%CI: 3.55–83.74; P b 0.001). The patients were 6.88 times more likely (95%CI 1.84–25.75; P = 0.004) to express “resignation” than their carers. The most frequent medical issues were related to: “medications”, “frequency of crises”, “work impact”, “quality of life”, “psychomedical integration of care”, and “development of new drugs”. After adjusting for severity in a logistic regression analysis, patients were more likely to express concerns for the frequency of crises than carers (OR 3.57, 95%CI: 1.16–11.04; P = 0.027). Conclusions: Patients' and carers' priorities, based on intense personal insight, represent a starting point to work for shared outcome measures in clinical trials and shared agenda in research, including research of strong evidence in complex intervention as service models for people with epilepsy. © 2017 Elsevier Inc. All rights reserved.
1. Introduction A sensible research agenda must consider the needs of patients and caregivers, alongside those of clinicians, researchers, decision makers, and even the pharmaceutical industry so that each stakeholder is aware of the others' needs [1]. Academic institutions have generally left the choice of topics and methodologies to their professional and scientific committees, often ignoring nonprofessionals when making these decisions. Patients' preferences should be included in the choice ⁎ Corresponding author at: Department of Neurophysiopathology, Azienda Ospedaliera di Perugia, Ospedale S. Maria della Misericordia, S. Andrea delle Fratte, 06156 Perugia, Italy. E-mail addresses: [email protected], neurofi[email protected] (M.G. Celani).
of topics and in the design of clinical trials, especially regarding the selection of outcome measures, the most relevant and important part of research for the end users of treatments and care [2]. Seizure frequency, in epilepsy research, is often the primary outcome in clinical intervention trials; however, it may not be the best predictor of patients' daily life skills, and it cannot be assumed that good psychosocial outcomes will be associated with the attainment of a seizure-free state. It may be necessary to go beyond seizure counts in the assessment of the disease and beyond the simple appearances of symptoms. Similarly, the obsessive research for a new drug may not be the exclusive aspiration of a patient. It could be useful to look for further and innumerable aspects that undermine the life of the ill individuals and their carer [3]. The aim of this work was to elicit perspectives, ideas, and values from people with epilepsy and their carers in order to offer clinicians
http://dx.doi.org/10.1016/j.yebeh.2017.08.028 1525-5050/© 2017 Elsevier Inc. All rights reserved.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
2
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
and researchers new indications, suggestions, and hints able to reshape the future of treatment and care in a joint effort. 2. Materials and methods 2.1. Inclusion criteria Adult patients affected by epilepsy, attending the outpatient service of the Neurophysiopathology Department, Santa Maria Misericordia Hospital, Perugia, Italy, and carers were invited to participate in this study. Patients had to be over 18 years of age, able to give informed consent, and have a clinical diagnosis of epilepsy. Caregivers, including all the people involved in the care of the patient: family members, friends, significant others, and personal home caretakers, had to be over 18 years of age and able to give informed consent. 2.2. Exclusion criteria Participants who were unable to provide informed consent or were affected by learning difficulties and medium to severe mental challenges were excluded.
2.3. Methodology Patients and carers were recruited by phone using the patients' lists from the outpatient epilepsy clinic. During recruitment, 15 phone call invitations were made by a researcher with the aim of having 6 to 8 participants in each group Fig. 1. During each phone call, the researcher informed the potential participant of the aim, methodology, and practical aspects of the study. During the conversation, an evaluation of the patient was made. Participants who agreed to take part in the study were divided into different groups based on the severity of his/her disease and were classified, using the different grades of the Modified Rankin Scale 0–2, 3, 4–5, into three different groups: mild, moderate, and severe. This scale is a simple, easy-to-administer tool, and can be given over the phone by both experts and nonmedical study personnel. It is used to measure functional outcome in a broad population of patients with neurological diseases [4–6]. Patients in the mild group are those who are self-sufficient, with a good pharmacological control of their seizures, and no loss of personal or social abilities. Moderate patients often require the assistance of a caretaker in order to accomplish some of their personal and social activities. Severe patients depend almost completely on a caregiver for every need. Once each participant with epilepsy was
Fig. 1. Flow chart patients' and carers' responses.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
categorized into the appropriate group, his/her caregiver (or the caregiver of another patient with the same characteristics) was included in the equivalent group. Two groups of patients and two groups of caregivers were formed for each level of severity of disease, except for the severe patients (one group only) for a total of 11 focus groups (Table 1). In order to obtain and measure patient-centered objectives, we used focus groups, a qualitative type of methodology used in health research [7]. There were 4 to 10 people in each focus group; all groups discussed the same topic, answering predetermined and semistructured questions (see Appendix 1 for focus group schedules), and each session was recorded. Two psychologists (RP and LB), trained in engaging in interactive group settings, conducted the recruitment, obtained verbal and written consensus to participate, and were the moderators of the focus groups. They guided the opening of the discussions and regulated the session so that all participants had the opportunity to actively participate. One moderator asked the questions and regulated response times, while another managed the recording, the moderators also helped to stay on topic, balance group participation to help make all group members feel comfortable. Focus groups averaged 90 min. At the end of each session, the moderator gave a short summary which included all the points of view that had emerged during discussion to make sure that each participant felt their contribution to the discussion that had been considered. The recorded group discussions were transcribed into text files. The transcripts were independently read by the moderators in order to blindly identify the key semantic meanings expressing the patients' perceived needs, as well as the emotions regarding the disease and its consequences. Two neurologists (MGC and TAC) also examined the files to highlight health and medical-related matters. After this first step, the findings were discussed to compare results and identify equivalent key themes from participants' reports. 2.4. Data analysis Each transcript was analyzed with “Concordance” software (version 3.3, University of Dundee, UK Watt R.J.C) using disease severity as classifier (mild, moderate, and severe), the subject's severity status (pt/c), the language used (formal language or dialect), and, for carers, the proximity of their relationship to the patient. Word frequency was computed after removing the usual stop words (articles, prepositions, etc.). The number of words by subject did include the stop words. For this analysis, only the following classifiers were used: severity and status. Needs, emotions, and medical-related matters were also classified for future semantic analyses; for this analysis, however, they were transformed into headwords and counted with software facilities. The results stratified by classifiers were then exported as “.xml files” and loaded into R and SPSS for statistical analysis. The descriptive analysis used mean ± (SD) Standard Deviation and/ or 95% Confidence Interval (95%CI) for normally distributed variables, median with 95%CI estimated by bootstrapping for nonnormally distributed variables, proportions with 95%CI as appropriate, for nominal variables. Comparisons made use of the t-test and analysis of variance (ANOVA) for normally distributed variables; Mann–Whitney and Kruskal–Wallis
3
tests for variables significantly deviating from normality, with the chi square test, the chi square test for trend, the exact Fisher test, and/or the logistic regression analysis for nominal variables. The ANOVA of variables significantly deviating from the assumption of normality was performed by bootstrapping using the unrestricted permutation of observations [8], rather than with the rank ANOVA. 3. Results We analyzed the responses of 25 patients and 36 carers, of whom 31 first-degree relatives (mother/father; sister/brother) and 5 spouses (Table 1). Women represented 50% of carers and 56% of patients (P = 0.644); gender distribution was not modified by severity (P = 0.537). As expected, age was significantly higher among carers (median: 65; range: 49–76) than patients (median: 40; range :19–76; P b 0.001, Mann–Whitney test) and this difference was homogeneous throughout severity levels (P = 0.469; Kruskal–Wallis test). The first item examined was whether the participants spoke about themselves. To this aim, we examined the frequency of the words related to “I” (I, me, my, mine, etc. [Table 2]). Being “a patient” determined a significantly greater tendency to speak about themselves, in comparison with being “a carer”. The severity of disease did not influence how often patients referred to themselves during the focus groups, and there was no evidence of interaction between severity and status. The model that excluded the interaction term showed a significant effect of the status (P b 0.001) and not of its severity (P = 0.323). 3.1. Analysis of classifiers 3.1.1. Propensity to express The predisposition of participants to express needs (status = 0.305; severity = 0.083), emotions (status = 0.138; severity = 0.089), or health-related issues and medical matters (status = 0.998; severity = 0.820) did not differ between pt/c or by severity of disease (data not shown). Monitoring instead the density of expression (average number of words uttered per need, emotion or medical issue expressed), patients' speech was significantly denser in needs and medical issues regardless of severity, while the density in emotions increased with severity, regardless of status (Table 3). 3.1.2. Expressed needs Overall, 12 different needs were expressed 169 times by the 61 participants (pt/c). More than 40% of the participants asked for “assistance” i.e., home health care, psychological support, a more intimate and committed patient–doctor relationship, better use of skilled social and health workers, and specific resources for patients with epilepsy. In terms of total number of expressions of needs, the most relevant ones were assistance, experience-sharing, knowledge, control, clarity, and security. If we assume that the “mean number of expressions per subject expressing” represents the “relevance for the individual”, their most relevant expressed needs were assistance, experience-sharing, knowledge, and outpouring. The subject expressing the word “assistance” repeated it more than 2.3 times (Table 4).
Table 1 Participants in focus group divided by status and disease severity. Disease severity by Modified Rankin Scale
Mild (0–2)
Moderate (3)
Severe (4–5)
Number of patients in each focus group
FOCUS n. 1:4 FOCUS n. 7:7 11 FOCUS n. 2:4 FOCUS n. 8:6 10
FOCUS n. 3:4 FOCUS n. 9:6 10 FOCUS n. 4:8 FOCUS n. 10:7 15
FOCUS n. 5:4 4 FOCUS n. 6:5 FOCUS n. 11:6 11
TOT Number of carer in each focus group TOT Numbers in bold are participants.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
4
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
Table 2 Mean percent proportion ± SD (N) of terms related to “I” on the total number of significant words pronounced. Mild
Moderate
Severe
Table 4 Distribution of expressed needs. Expressed needs
n. of subjects expressing (% of total subjects)
n. of expressions (% of total expressions)
Mean ± SD n. of expressions per subject expressing
Assistance Experience-sharing Knowledge Clarity Control Planning Security Comprehension Predictability Outpouring Doing Respect Total
26 (42.6%) 14 (23.0%) 12 (19.7%) 9 (14.8%) 6 (9.8%) 6 (9.8%) 5 (8.2%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 2 (3.3%) 1 (1.6%) 61
60 (35.5%) 32 (18.9%) 25 (14.8%) 10 (5.9%) 11 (6.5%) 7 (4.1%) 9 (5.3%) 3 (1.8%) 3 (1.8%) 6 (3.6%) 2 (1.2%) 1 (0.6%) 169
2.31 ± 1.57 2.29 ± 2.20 2.08 ± 1.08 1.11 ± 0.33 1.83 ± 0.75 1.17 ± 0.41 1.80 ± 0.84 1.00 ± 0.00 1.00 ± 0.00 2.00 ± 1.00 1.00 ± 0.00 1.00 ± − 1.88
Total
Carers 2.49 ± 1.32 (10) 2.64 ± 0.91 (15) 2.76 ± 1.92 (11) 2.64 ± 1.36 (36) Patients 5.62 ± 2.26 (11) 4.16 ± 1.13 (10) 5.80 ± 0.48 (4) 5.06 ± 1.79 (25) Total 4.13 ± 2.43 (21) 3.25 ± 1.24 (25) 3.57 ± 2.15 (15)
More patients than carers expressed the need for “experiencesharing” (40.0%, 95%CI: 21.1%–61.3% vs. 11.1% 95%CI: 3.1%–26.1%; Fisher test P = 0.013) and “control” (20.0%, 95%CI: 6.8%–40.7% vs. 2.8%, 95%CI: 0.0%–14.5%; P = 0.038). No other significant difference was seen for the other needs expressed. To estimate the effect of status (pt/c) and severity of disease on the distribution of the most frequently expressed needs, logistic regression analysis was applied. The only significant effect was seen by status (pt/c) on experience-sharing (P = 0.006). After adjusting for severity, patients expressed the need for experience-sharing more than their carers (OR 7.29; 95%CI: 1.76%–30.18%). If we consider patients and carers together, stratified by disease severity, the need for planning (P = 0.010; chi square for trend) and for assistance (P = 0.227, NS) was directly related to disease severity while the need for control was inversely correlated (P = 0.054, NS) (Fig. 2). In other words, the need for control was more relevant for people with a mild disease while, as expected, the need for planning and assistance was important in people with more disabilities. 3.1.3. Expressed emotions Overall, 20 different emotions were expressed 341 times by the 61 participants (pt/c). More than 50% of the participants expressed “hope”, i.e., faith, having a normal life, more effective medications, and decrease in seizure frequency. In terms of total number of expressed emotions, the most relevant were hope, anger, fear, and resignation. In terms of relevance for each participant, anger, resignation, fear, anxiety/apprehension, acceptance, hope, and shame ranged from 2.36 to 1.56 (Table 5). The frequency of participants expressing “resignation” at least once was greater among patients than carers (respectively 48.0%, 95%CI: 27.8%–68.7% vs. 13.9%, 95%CI: 4.7%–29.5%; Fisher test P = 0.008). The frequency of “disappointment” followed a similar trend (12.0%, 95%CI: 2.5%–31.2% vs. 0.0%, 95%CI: 0.0%–9.7%; P = 0.064). On the contrary, the frequency of participants expressing “anger” was much greater among carers than patients (61.1%, 95%CI: 43.5%–76.9% vs. 12.0%,
95%CI: 2.5%–31.2%; P b 0.001). No other significant difference was seen for the other emotions expressed. The distribution of the most frequently expressed emotions was tested by logistic regression analysis to estimate the effect of status and severity. The only significant effects were by status on “anger” and on “resignation”. After adjusting for severity, carers were more likely to express anger, in comparison with patients (OR 17.23, 95%CI 3.55–83.74; P b 0.001). Patients were 6.88 times more likely (95%CI 1.84–25.75; P = 0.004) to express “resignation” than their carers. Considering the frequency of repetition per subject, there was no particular emotion that was repeated more frequently. When stratified by disease severity, the proportion of participants (pt/c) expressing shame was borderline associated with increased severity (P = 0.051), while in terms of proportion of responses made by participants, the relevance of “hope” (P = 0.002 chi square for trend) and “shame” (P = 0.002) highly increased, while “acceptance” (P = 0.003) decreased with severity of disease (Fig. 3).
3.1.4. Health-related issues and medical matters — physicians' perspectives Overall, 25 different medical topics were expressed 358 times by the 61 participants. The most frequent ones were related to “medications”, “frequency of crises”, “work impact”, “quality of life”, “psychomedical integration of care”, and “development of new drugs”. Among the total number of expressed topics, the most relevant ones were “work impact”, “medications”, “new drugs”, and “drug adverse reactions” (Table 6).
Table 3 Median number of words (±95%CI) pronounced per need, emotion or medical issues expresseda. Item Needs
Emotions
Medical issues
a b c
Carers Patients Total Pb Pc Carers Patients Total Pb Pc Carers Patients Total Pb Pc
456 [206–1261] 565 [440–759] 213 [144–321] 297 [231–393] 240 [185–451] 440 [310–565] 0.09195 Status = 2e−04; severity = 0.786; interaction = 0.832 146 [93–207] 232 [217–257] 102 [71–120] 231 [126–306] 107 [89–145] 232 [197–270] 0.00071 Status = 0.1486; severity = 0.010; interaction = 0.679 78 [49–176] 78 [64–144] 40 [27–67] 51 [37–85] 52 [40–78] 68 [56–90] 0.27204 Status = 0.009; severity = 0.783; interaction = 0.134
413 [208–694] 116 [88–148] 250 [133–416]
456 [378–630] 213 [148–310]
0.00025
139 [76–416] 130 [99–177] 139 [99–189]
196 [152–228] 120 [99–155]
0.08754
63 [33–187] 40 [28–216] 47 [33–123]
74 [63–119] 46 [38–56]
0.00113
Mann–Whitney test for status. Kruskal–Wallis test for severity. ANOVA (bootstrapping, 5000 replications).
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
5
Fig. 2. Proportion of subject expressing a need for planning assistance and control by disease severity.
Fifty-six percent of the patients (95%CI: 34.9%–75.6%) expressed concern for drugs at least once compared with 22.2% of the carers (95%CI 10.1%–2801% Fisher exact test P = 0.014); “concern to adjustment of new lifestyle” 28.0% (95%CI 12.1%–49.4%) vs. 5.6% (95%CI 0.7%–18.7%) P = 0.024, and “fear of seizure recurrence” 52.0%(95%CI 31.3%–72.2%) vs. 22.2% (95%CI 10.1%–39.2%) P = 0.025. On the contrary, carers more frequently than patients expressed greater “concern for work” (44.4%; 95%CI 27.9%–61.9% vs. 20.0%; 95%CI 6.8%–40.7%; P = 0.059 borderline to significance). No other significant differences were seen. The distribution of the issues expressed more frequently (medications, frequency of crises, work, and quality of life) was also tested by using logistic regression analysis, in order to estimate the effect of status (pt/c) and severity adjusted for each category. The only statistically significant effect was for patients who were more likely to express concerns for “fear of seizure recurrence” than carers (OR 3.57; 95%CI 1.16–11.04; P = 0.027). No other significant differences were seen.
Table 5 Distribution of expressed emotions. Emotion
n. of participants expressing (% of total participants)
n. of expressions (% of total expressions)
Mean ± SD n. of expressions per subject expressing
Hope Anger Fear Sorrow Resignation Acceptance Anxiety/Apprehension Shame Bewilderment Guilt Disappointment Detachment Remorse Happiness/Serenity Reject Curiosity Gratitude Exasperation Surprise Sarcasm Total
33 (54.1%) 25 (41.0%) 30 (49.2%) 22 (36.1%) 17 (27.9%) 17 (27.9%) 14 (23.0%) 9 (14.8%) 7 (11.5%) 4 (6.6%) 3 (4.9%) 2 (3.3%) 1 (1.6%) 2 (3.3%) 2 (3.3%) 1 (1.6%) 1 (1.6%) 1 (1.6%) 1 (1.6%) 1 (1.6%) 61
60 (17.6%) 59 (17.3%) 59 (17.3%) 27 (7.9%) 35 (10.3%) 32 (9.4%) 27 (7.9%) 14 (4.1%) 8 (2.3%) 4 (1.2%) 3 (0.9%) 2 (0.6%) 2 (0.6%) 2 (0.6%) 2 (0.6%) 1 (0.3%) 1 (0.3%) 1 (0.3%) 1 (0.3%) 1 (0.3%) 341
1.82 ± 1.69 2.36 ± 2.00 1.97 ± 1.35 1.23 ± 0.43 2.06 ± 1.60 1.88 ± 1.17 1.93 ± 1.94 1.56 ± 0.53 1.14 ± 0.38 1.00 ± 0.00 1.00 ± 0.00 1.00 ± 0.00 2.00 ± − 1.00 ± 0.00 1.00 ± 0.00 1.00 ± − 1.00 ± − 1.00 ± − 1.00 ± − 1.00 ± − 1.77
4. Discussion The real novelty of our study is that by using the text analysis technique we turned qualitative data into semiquantitative ones. Two professional blinded experts indexed the content of focus groups, separately coded important terms, and categorized open-ended common responses. Concordance between the two coders was 90%. By doing so, we reduced the gap between what people actually said in focus groups and what the moderator interpreted in order to minimize interpretative biases. We chose, as the primary tool of our research, focus groups' discussions, in order to freely elicit how people feel, function, share, and compare their different points of views instead of the more predetermined self-rated questionnaires. We grouped participants by severity of disease and by their status, i.e., being a patient or a carer, with the aim to understand the real differences in perspectives between patients and family members and the way these perspectives change depending on the severity of the disease. We confirmed our hypothesis that there is a large difference in perspective between patients and carers. The amount of words related to “I” (I, me, my, mine) that we coded as “self-expressed” suggests that patients speak of and for themselves, whereas carers speak on behalf of the patients they care for. We speculate that this peculiar result depends on a cultural and religious attitude of moral awareness (strong sense of duty) and on the ethical responsibility in looking after people [9]; for this reason, we cannot exclude that this feeling is particularly strong in Italy where there is a historical imprinting of Catholicism and might reflect the small and homogeneous sample size. Furthermore, patients tend to go straight to the issue with few words to express a need or medical aspect, while carers need significantly more. Another aspect of our study is about what we refer to as “needs”. The most relevant “needs” that emerged from all participants were “assistance”, “experience-sharing”, “knowledge”, “clarity”, “control”, and “planning”. “Assistance”, considered as the access to tailored and more dedicated health services, is the most specific need expressed by all participants, while “experience-sharing” and “knowledge” are unique needs expressed particularly by patients. People affected with the disease wish to have the opportunity to compare with other patients, support each other, and have a deeper knowledge of the medical aspects of epilepsy in order to have better control of the disease and medications, and consequently, reduce fear. In this study, the need defined as “planning” is considered the necessity of having easier access to employment possibilities and social life, from which patients with epilepsy are often excluded. “Planning” is strongly correlated to severity of disease, while “Control”, intended as the possibility to manage the different aspects of the disease, like related risk of seizures, medication
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
6
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
Fig. 3. Statements expressing hope, shame, and acceptance estimated as a proportion of the total statements expressing an emotion stratified by disease severity.
side effects, the unpredictability of the seizure event, and controlling families and carers, is inversely proportional to severity. We also explored participants' “emotions”, in relation to the disease. In this regard, patients expressed resignation and failure (i.e., “I have accepted it”, “my life had to slow down”, “I had to give up professional soccer”) while carers expressed mostly anger (“the whole family and our routine was effected by the diagnosis, and the relationship with the ill person is often hard because his/hers life is ruined, and so is ours”). Surprisingly, alongside the negative emotions, participants expressed “hope”, wishing for a “normal” life, more effective medications, and decrease of seizure frequency. We found a marginal statistically significant emotion regarding “shame” among severely affected people, and we were consequently able to speculate that there is still the perception among patients of being rejected and most of all stigmatized by society [10]. Stigma still seems to be one of the most hurtful and persistent issues. Kale more than 20 years ago stated that “the history of epilepsy can be summarized as 4000 years of ignorance, superstition, and stigma, followed by
100 years of knowledge, superstition, and stigma”. While conducting our study, we came across the fact that very little has changed regarding this disease, despite the numerous initiatives, including the famous joint project called “Out of shadow”, started in 1997 [11]. Stigmatization leads to discrimination, and people affected by epilepsy have been more often than not targeted throughout history and cultures with prejudicial behaviors. When the two physicians interpreted the focus group texts, they deduced that approximately 20% of the total expressions by participants were about epileptic seizures (number, frequency, length) while 24.6% were about medication (drugs in general, new medications, adverse events). A concerned spouse and caregiver stated that because of the psychoactive effects of the medication, he no longer had any idea of who his wife actually was. A straightforward consequence, given the effective cognitive behavioral effects of medications and the related concerns of consumers, is the necessity to conduct prolonged follow-ups throughout the years in order to observe psychological variations of personality. 4.1. Limitations
Table 6 Distribution of expressed medical issues. Medical issue
No. subjects expressing (% of total subjects)
No. expression (% of total expressions)
Drugs Frequency of crises in general Work Quality of life Psych medical cure integration New drugs Medical assistance Small crises intensity Crises intensity Adverse reactions Hobbies Proper information Driving license Frequency of severe crises Education to new lifestyle Frequency of small crises Social interaction Genitoriality Drug dependence/need Psychiatric comorbidity Drug efficacy without ADRs Surgery Individualized project Intercritic time Drug cost Total
22 (36.1%) 21 (34.4%) 21 (34.4%) 17 (27.9%) 16 (26.2%) 16 (26.2%) 15 (24.6%) 14 (23.0%) 13 (21.3%) 12 (19.7%) 12 (19.7%) 12 (19.7%) 12 (19.7%) 10 (16.4%) 9 (14.8%) 9 (14.8%) 9 (14.8%) 5 (8.2%) 5 (8.2%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 3 (4.9%) 1 (1.6%) 61
27 (7.5%) 24 (6.7%) 31 (8.7%) 24 (6.7%) 23 (6.4%) 25 (7.0%) 22 (6.1%) 16 (4.5%) 17 (4.7%) 25 (7.0%) 15 (4.2%) 16 (4.5%) 15 (4.2%) 10 (2.8%) 10 (2.8%) 9 (2.5%) 14 (3.9%) 6 (1.7%) 7 (2.0%) 4 (1.1%) 4 (1.1%) 4 (1.1%) 6 (1.7%) 3 (0.8%) 1 (0.3%) 358
This study was carried out among a population with epilepsy attending the outpatient clinic of a tertiary epilepsy center in a single Italian region. Therefore, it may reflect the cultural and the social aspects of this geographical area, along with the characteristics of the local healthcare system and the environment where the participants live. We have taken into account that some aspects of our research may not apply to different contexts, even with a similar reality. Results could differ even more when compared to care and disease management in developing countries where the “organization of care” must be adapted to the existing healthcare system [12]. The major limitations of this study are the small sample size and the fact that participation to the focus groups was fully voluntary. People attending the focus groups were about one-third of those contacted. Taking into account that participation required time and effort, it is likely that the participants were those most motivated, consequently, these results might not be extendable to the overall population. 4.2. Strengths This is one of the few studies that analyze the precious information that comes from listening to people who carry the burden of disease. The analysis was carried out using a semiqualitative method to guarantee that the expressions used cannot be manipulated. This perspective offers the possibility of bridging the gap that exists between people who receive funding and those who make the decision regarding future
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028
M.G. Celani et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
research and what patients, families, and the public would like to see explored. This would help us to move toward patient-centered medicine and tailored outcome measures, implemented in a patientcentered health organization. 5. Conclusions The analysis of the focus groups shows that reduction of seizure frequency, the oldest and the most common outcome measure in epilepsy clinical trials, is still the most important for people affected by the disease. On the other hand, drug side effects, which are at best considered a secondary outcome from the main study objectives in clinical trials, become extremely important to daily living both for patients and carers. It is for this reason that we feel a primary outcome in epilepsy research that should include a sensible balance between benefits and risks at short and long term. A second aspect can be summarized with the comprehensive care concept, derived from expressed patients' and carers' needs. Patients have the desire to become empowered and knowledgeable about the correct health behaviors that need to be learned in order to better control their seizures, to better adapt themselves to the constraints of their disease and promote their well-being. In addition, people affected by the disease expect that health services will include highly specialized epilepsy care in liaison with multidisciplinary teams available in their communities. Service models for epilepsy have been evaluated in different clinical trials and synthesized in systematic reviews published in the Cochrane Library [13]. The reviews conclude that, at present, there is some evidence of effectiveness when a specialist epilepsy nurse is at hand, and self-management education is offered. Evidence on other more complex intervention types is still unclear and should be better investigated and consequently proposed to policy makers. The patients' and carers' voices, elicited in our study, drive us to organize the future agenda in epilepsy research in this direction [14]. Conflict of interest This study had no financial support and authors have no conflict of interest. Appendix 1 Topic area for focus group interviews: 1. 2. 3. 4.
Recall the first time you were ill. What did you feel when you were told you had the disease? What problems did you have to face then and now? What do you consider the most important aspects to be improved in disease care?
Appendix 2 Procedure used to classify focus groups' results 1. Each focus group was audio recorded and transcribed into a text file. 2. Each speaker was code-identified, classified by status (patient/ carer) and severity (mild/moderate/severe). 3. All transcripts were crosschecked against the recordings independently by the two moderators (the two psychologists).
7
4. Any discrepancy between transcripts was resolved by agreement between the moderators. 5. Any reference to personal identity was removed from all text files. 6. Copies of each text file with markers for status and severity were distributed to the two psychologists and the two neurologists. 7. Each reader independently marked and assigned an emotion, need, or medical issue to the individual statements in the text files. These could be part of a sentence, a whole sentence, or span over more than one sentence. A statement included one complete semantic content. 8. The readers compared each individual statement and the semantic content applied; differences were resolved by agreement. 9. The two psychologists individually marked every statement expressing any single emotion, need, or both. The two neurologists individually marked every statement expressing a health-related issue and/or medical matter. 10. Following a brief discussion, discrepancies were resolved without the use of external support as foreseen in the protocol. 11. All classifications were revised jointly by the two psychologists for needs and emotions, and by the two neurologists for health issues and medical matters. Overlapping or very similar classifications were reexamined to determine whether they could be compounded into a single classification. Reclassification and compounding was performed by agreement between readers. 12. Final classifications were inserted as markers into the transcripts. For this study, the markers were changed into headwords and elaborated using the Concordance software, they were classified by speaker, severity, and status. The version containing all markers was archived for future semantic analyses. 13. Results from Concordance were loaded into statistical software (R and SPSS) to perform the planned analyses. References [1] Liberati A. Need to realign patient-oriented and commercial and academic research. Lancet 2011;378(9805):1777–8. [2] Crowe S, Giles C. Making patient relevant clinical research a reality. BMJ 2016;355: i6627. [3] Heath I. How medicine has exploited rationality at the expense of humanity: an essay by Iona Heath. BMJ 2016;355:i5705. [4] Candelise L, Pinardi G, Aritzu E. Interview for stroke outcome assessment. Cerebrovasc Dis 1994;4(5):341–3. [5] Patel N, Rao VA, Heilman-Espinoza ER. Simple and reliable determination of the modified Rankin scale score in neurosurgical and neurological patients: the mRS9Q. Neurosurgery 2012;71(5):971–5 [discussion 975]. [6] Savio K, Della Pietra GL, Oddone E. Reliability of the modified Rankin scale applied by telephone. Neurol Int 19 2013;5(1):e2. [7] Lehoux P, Poland B, Daudelin G. Focus group research and “the patient's view”. Soc Sci Med 2006;63:2091–104. [8] Chapman & Hall. In: BFJ Manly, editor. Randomization, bootstrap, and Monte Carlo methods in biology3rd ed. ; 2007 [London]. [9] Kleinman A. Caregiving as moral experience. Lancet 2012;380(9853):1550–1. [10] De Boer HM, Mula M, Sander JW. The global burden and stigma of epilepsy. Epilepsy Behav 2008;12(4):540–6. [11] Kale R. Bringing epilepsy out of the shadows. BMJ 1997;315:2–3. [12] Birbeck GL. Epilepsy care in developing countries: part I of II. Epilepsy Curr 2010; 10(4):75–9. [13] Bradley P, Bruce L, Nigel F. Care delivery and self-management strategies for adults with epilepsy. Cochrane database of systematic reviews. The Cochrane Collaboration; 2016 [Issue 2. Art. No: CD006244]. [14] Lynch S. Measures need to capture patients' views and experiences more effectively. BMJ 2013;346:f1553.
Please cite this article as: Celani MG, et al, Is it worth it, or possible, to measure what matters to patients with epilepsy and their caregivers? Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.08.028