**4. Discussion**

In the UK, treating physicians were not identified and therefore could not be interviewed. Similarly, in Italy recruitment was via a patients' organization and did not involve treating physicians. In Latvia, participant selection was based on ICD-10 diagnoses. This was following advice from the country's only secondary referral center for ME/CFS. The diversity of recruitment methods in the three countries was probably a source of strength rather than weakness, as the findings from the survey from the three countries revealed some very similar problems and concerns, suggesting that patients' experiences were universal in nature, and not confined to any one country or health care system. It is interesting that, despite differences in health care systems, diagnostic methods, recruitment methods and survey media, there was a very considerable similarity in the experiences of respondents in all three countries.

Respondents in all three countries were similar in terms of average age and had the same preponderance of females. There were no significant differences in the proportions of respondents who were married or living alone, or who had post-school educational qualifications. In Latvia, the UK, and Italy, net household incomes were lower than the national average, indicating the impact of illness on incomes in all countries. As regards illness managemen<sup>t</sup> and progression, there were significant differences between the three countries. Italians reported more symptoms, more investigations and more treatments than respondents from the other countries, and the Latvians the least. This may be due to the absence of appropriate guidelines for the managemen<sup>t</sup> of the disease in Italy. We did not elicit any information on the nature of the investigations carried out, because the purpose of this question was to obtain a measure of the extent to which doctors in the three countries were taking seriously their patients' illnesses and were actively working to investigate them.

Symptom fluctuation was significantly more marked among the UK respondents than among the others. It was noteworthy that graded exercise therapy, in the UK, was found to be universally ineffective, and none of the Italians reported having had this therapy. There were reported differences between the three countries in who had responsibility for providing medical care, but these may reflect differences in the managemen<sup>t</sup> of the disease in each country. Thus, GPs more frequently had principal responsibility for medical care in Latvia than in Italy or the UK and this probably reflects the fact that in Latvia

GPs perform the gate-keeper role for patients in the diagnostic and treatment process. Healthcare professionals other than doctors were more frequently involved in clinical care in the UK than in either Latvia or Italy, which is likely to be related to the pattern of delivery of primary care in the UK via the National Health Service.

In terms of personal care also, differences in response between the three countries were more likely to reflect differences in the way in which social care is delivered in the three countries. Thus, non-family care assistance was almost entirely confined to the UK, where capacity for self-care was less prevalent that in the other countries. Similarly, variations in difficulty explaining the disease to physicians was least widespread in Latvia, but this may be attributable to the fact that physicians there were centrally involved in the identification of potential respondents. Other variations, e.g., in explaining the illness to family, friends or employers, may reflect differences in society in the three countries. Italians were most likely to have trouble in explaining the illness to families, while Latvians had the least difficulty explaining the illness to friends or employers, which suggests that there may be greater understanding of the illness in the general population than is the case in either Italy or the UK.

While there were differences between the three countries in perceived quality of life both before and during illness, the trend was similar in Italy, Latvia, and the UK, with marked diminution in quality of life being reported in all three countries as a result of illness. In Latvia, the smaller gap in quality of life as a result of illness could be explained by a greater decrease in quality of life before diagnosis, with a subsequent smaller decrease in quality of life after diagnosis and initiation of treatment. While we do not wish to overinterpret these data, it is likely that the initial differences between the three countries reported prior to illness, for example in income and educational attainments, may reflect overall socioeconomic differences between them, while there is substantial convergence between all three countries in patient experience once illness becomes established.

The strength of this study is that this is the first study of ME/CFS patients conducted on a transnational, comparative basis in Europe. It demonstrates that the basic demographic features of the illness, in terms for example of the average age of participants and the gender distribution, are very similar in the three countries studied. Where the responses from the three countries differ, this largely reflects socioeconomic differences between countries, or differences in the way in which medical services are delivered. Although there was not a single template for the recruitment of patients, we endeavored to ensure the comparability of patients through the recruitment process. Thus, in Italy, only patients with a recognized diagnosis of CFS/ME were selected. Similarly, in Latvia, from almost 300 respondents to the on-line survey on CFS symptoms, only participants with ICD-10 diagnosis codes G93.3, R53 and B94.8 were involved in data analysis for the purposes of this paper, while in the UK patients with medically confirmed diagnoses of ME/CFS were self-selected via an internet-based patient support group.

The finding that ME/CFS has a substantial impact on net household income is consistent with the previous conclusion of the socioeconomics working group of EUROMENE that the economic impact of ME/CFS is substantial [5]. The reports of household income are worrying, given average 2018 incomes of €25,642 in the UK, €19,208 in Italy, and €8740 in Latvia [12]. International comparisons are very difficult due to the heterogeneity of average incomes in European countries, In Italy, 44.1% of respondents declared an income higher than €15,000, while the comparable proportion among UK respondents was 54.6% of respondents compared with 92% of the UK general population in 2017–18 [14]. In Latvia, annual incomes among the general population are lower than in Italy and the UK. No Latvian respondents reported an annual income per household member in excess of €15,000. Despite these differences between the three countries under consideration, the pattern is consistent; in all three countries, the respondents tended to report net incomes per household member which were substantially lower than those found among the general population. This underlines the negative impact of having ME/CFS on individual productivity and capacity to work and indicates that this impact is substantial. Our findings

are consistent with other research demonstrating a very substantial impact of ME/CFS on productivity costs. Thus, Reynolds et al. analyzed data from a surveillance study of ME/CFS in Wichita, Kansas, and concluded that lost productivity due to ME/CFS was substantial both in absolute terms and in comparison with other major illnesses [15], while Collin et al. analyzed data from the UK CFS/ME National Outcomes Database, and concluded that ME/CFS causes huge productivity costs amongs<sup>t</sup> the small fraction of adults with ME/CFS who access specialist services [16].

The paper did not address the relationship between severity and economic impacts, the importance of which we emphasized in a previous paper [5]. It is likely, though, that productivity costs were higher among the more severely ill patients, because being housebound or bedbound, severely ill patients are generally unable to work at all. Health care system costs are more complicated, because many severely ill patients receive no support or help from the health care system at all, due to the failure of primary care physicians to diagnose the illness. Such failure is widespread, with evidence that between a third to a half of all GPs, over several decades and a variety of geographical locations, expressing disbelief or failing to recognize ME/CFS as a genuine clinical entity [17,18].

This is consistent with the finding that a large proportion of respondents, particularly in Italy and the UK, have difficulty explaining their illness to doctors, as reported by the working group's literature review of knowledge and understanding of ME/CFS among GPs [17], which established that disbelief and lack of knowledge were widespread in primary care, while our survey of perceptions of GP knowledge and understanding among EUROMENE participants suggests that this problem exists throughout Europe [18].

This study initiative aimed to identify policy measures designed to alleviate the burden of disease on patients and their families, and on governments, in particular by reducing delays in diagnosis, and has enabled certain changes in the way in which services are delivered to be identified. Thus, in Latvia, the survey was paralleled by a survey of GPs, in which a number of improvements were suggested, including establishment of a consortium of specialists, the creation and use of clinical algorithms and patient pathways, better information for patients, reimbursement from public funds of psychotherapists' consultation fees, additional training, and more time for patient consultations. Another possible improvement recommended from Latvia was the establishment of a disease register, which would facilitate disease managemen<sup>t</sup> by GPs, the development of patient pathways, and improved disease monitoring.
