The aim of this study was to describe HRQOL in the general Swedish population using the FACT-GP instrument and a single item question concerning SRH. To our knowledge this is the largest sample (n = 2791) using the FACT-GP instrument and the results could hopefully serve as a normative benchmark to which other HRQOL-data could be given meaning.
Studies have shown that HRQOL varies between countries and cultures, suggesting an important need to provide normative data from different cultures (16, 17). Few studies have nonetheless described HRQOL in a general population, using either FACT-G or FACT-GP. These existing studies are based on data from US (12), Austria (4) and Australia (7). The study by Brucker et al (12) was performed on 1075 participants and compared FACT-G between a general population and a cancer population. The studies by Holzner and colleagues (4) and by Janda and colleagues (7) was based on 968 and 2727 participants respectively. Both last-mentioned studies compared FACT-GP scores with sociodemographic data.
HRQOL. The present Swedish population scored lower on FACT-GP in comparison with the other studies using FACT-GP to assess HRQOL (M 80.1 (12), M 85.9 (7) and M 86.5 (4)). This may be surprising since the present sample is skewed towards a higher socioeconomic status which would rather indicate a higher HRQOL score. There is no obvious explanation for the observed differences in FACT-GP score between the studied countries’ general populations. When comparing the present results to those of a study investigating different countries’ reference values for the HRQOL-instrument EORTC QLQ-C30, that result is in line with the present study indicating a lower global health/QoL scores for the Swedish population as compared to the Austrian (5). Though the results were not in line with the present when showing higher global health/QoL scores for a US population as compared to the Swedish (5).
Cancer. The results showed that there was no significant difference between participants who had been diagnosed with cancer or not in HRQOL as assessed by FACT-GP. The study by Holzner and colleagues (4) showed inconclusive results concerning HRQOL for patients with a previous cancer diagnosis in complete remission in comparison to the general population. That study reported similar FACT-GP scores in comparison to the general population for patients with previous breast cancer, but lower FACT-GP scores for patients with previous bone-marrow transplantation and higher FACT-GP scores for patients with Hodgkin’s disease (4). In the study of Janda (7) it appears that there were no difference in FACT-GP score between participants ever diagnosed with cancer or not (M = 83.9 SD = 15.6 vs M = 86.2 SD = 15.0), however this was lower in the present study (M = 78.78, SD = 18.81). One difference between the studies is the prevalence of cancer, in the study by Janda (7) 3% of the sample reported a previous cancer diagnosis as compared to 7.5% in the present study. No details concerning type of cancer and treatments given are available in either of the studies, making any deeper comparisons impossible. There is a trend towards increasing HRQOL in cancer patients over time. This pattern has been shown previously for example in malignant melanoma, where the greatest decline in HRQOL is observed at the time of diagnosis and the immediate post-treatment period, but after that, HRQOL slowly improves over time (18, 19).
SRH. To our knowledge, no previous study has investigated the correlation between SRH and FACT-GP, and the present results show a strong correlation that is robust even when controlling for age, gender, education and income. Though not studying SRH specifically, Holzner et al. [9] showed that participants suffering from chronic illnesses reported lower HRQOL on all subscales and overall FACT-GP as compared to healthy subjects. The study by Janda et al (7) investigated the effect of co-morbidities and found that HRQOL decreased with increasing number of co-morbidities for all subscales and also the overall FACT-GP score. The strong positive association between SRH and HRQOL are primarily related to better health outcomes (20, 21) and it is believed that health outcomes are associated with healthy behaviours like regular physical activity, less sedentary time, a healthier diet and non-smoking habits (22). When interpreting SRH elicited from a general population specially in relation to a disease -specific population, it has in many studies been noted that patients tend to rate their health to be higher than members of the general population, thus indicating that patients rate their own health as better than members of the general public do (23, 24). In effect, the importance of considering the values of both patients and those of the general populations when interpreting HRQOL- and SHR has been argued by several researchers (23, 25).
Gender. The present results indicated a statistically significant difference between men and women in the overall FACT-GP score. Men scored higher on each subscale except for SWB where women scored significantly higher. The only non-significant difference between men and women were for the subscale of FWB. A lower overall reported HRQOL by females was also seen in the Austrian study (4), were women reported significantly lower QOL values than men in the subscales PWB and EWB. A suggested explanation for the differences in that study was that males were potentially less inclined to admit an impaired QOL as compared to women (4). A similar difference between gender have been observed when using also other HRQOL-instruments (26, 27), it should however be mentioned that a previous study investigating HRQOL using FACT-GP observed no gender difference (7).
Age. The present results show that younger (< 25 year) and older age groups (> 45 years) tend to score higher in general. The observed results are not coherent with previous FACT-GP results, where the Austrian sample showed a trend of declining HRQOL with increasing age where the oldest age group (> 70 years) had significantly lower HRQOL than the younger ages (4). The study investigating FACT-G scores in the Australian sample did not find any significant differences between different age groups concerning HRQOL (7). Though, the present results are in line with a recently published study investigating country reference values for HRQOL using a the HRQOL instrument EORTC QLQ-C30 (5). The study showed higher global health/QoL scores for younger ages (18–29 years) as well as for higher ages (≥60 years for females and ≥70 years for males) as compared to middle ages (30–59 years) for both genders (5).
Income. The present results showed a trend towards an increasing HRQOL with increasing income, and to our knowledge no other study has investigated this correlation using FACT-GP. Previous studies have investigated if there is a correlation between occupational status and HRQOL and in the Austrian study (4), the participants’ occupational status did not have an effect on any of the FACT-GP subscales. In contrast, the participants working full-time reported higher HRQOL on FWB and overall summary FACT-GP score as compared to participants working part-time in the Australian study (7). Other studies, have shown that wealth and SRH have a strong positive correlation (28) and that higher income is more often associated with a positive evaluation of life, including well-being and life satisfaction (6).
Education. The present results indicated that participants who reported a higher level of completed education had a significantly higher overall summary FACT-GP score, and the same results were observed in both the Austrian (4) and the Australian study (7). The observed results are probably explained by the fact that more highly educated people are likely to have more satisfying work situations and generally to have more privileged lifestyles (4).
Birthplace. The present results did not show any significant differences between groups when analysing HRQOL-scores depending on birthplace. Thus, the present study indicate that overall FACT-GP score are not affected by ethnicity. It is not known if our results are in line with other studies since, to our knowledge, no other study have investigated the correlation between ethnicity by investigating location of birth in relation to current country of residency. In general, studies using other instruments than FACT-GP show inconclusive results concerning ethnicity and HRQOL (29, 30).
Living area. The present results did not show any significant differences between groups when analysing HRQOL-scores depending on living area like countryside, small or large towns or abroad. The results concerning living area are in line with the previous study with Australian participants showing no effect on HRQOL (7).
Strengths and Limitations. The main limitation of this study is that the participants had a somewhat higher age, educational level and income compared to the Swedish general population (15), thus indicating a risk for selection bias. As can be seen in Table 1, the sample is representative compared to the Swedish population regarding age and gender, albeit the youngest and oldest age groups are slightly underrepresented in our sample. The sample is less representative concerning educational attainment especially in the range of lower level education. Mean income in Sweden 2017 was approximately 294’ SEK and the median income was 270’ SEK (31). The most common reported income level in our data was a monthly salary of between 30 000 and 36 999 SEK. The sample thus show a higher level of socioeconomic status compared to the Swedish population in general. People with lower living standards are generally hard to reach with surveys even when a randomized postal survey with incentives are employed (32). To our knowledge, the web panel at the University of Gothenburg is still one of the most representative web panels in Sweden and may in some aspects provide a more representative sample compared to traditional postal or telephone surveys. Other limitations are that the present study is a cross-sectional study and that any conclusions concerning causality cannot be drawn, and that there are most probably several confounding factors that are not accounted for in the analysis regarding predictive validity to sociodemographic factors and SRH.