Translation
Back-translation of the IPSS (Hong Kong Chinese version 1) revealed that all questions, aside from the item on nocturia, and seven response options were not equivalent to the original English version The IPSS (Hong Kong Chinese version 1) uses a dichotomous leading question “Have you often…” which asks for a “Yes” or “No” response, whereas, the original English IPSS uses “How
often have you…” which asks about the frequency of the urinary symptom. The quality of life item was also problematic in the IPSS (Hong Kong Chinese version 1) as the translation did not incorporate the meaning “the rest of your life with your urinary condition”. Two of the response options for the symptom-related questions, and all five response options of the quality of life question were not equivalent to the original English version. Additional file
2 shows all non-equivalent questions and response options.
Through panel review, all nonequivalent questions and response options were modified to enhance their translational equivalence to the original English version. The panel also further refined one of the response options for the quality of life question (“unhappy”) even though back-translation was equivalent to original English version. This was done to enhance clarity and make it easier for respondents to compare the response options.
Table
2 shows the distribution of the responses for each item of the IPSS HKv2 by gender. For the total symptom score, 1.8% of subjects (2.4% female; 1.0% male) had the lowest possible score whilst no ceiling effect was observed. For the quality of life score, 4.7% of subjects (4.8% female; 4.7% male) had lowest possible score while 4.3% of subjects (3.2% female; 5.7% male) had highest possible score.
Table 2
Score distribution, corrected item-total correlation and reliability of the IPSS (HK v2) by gender
1. Incomplete emptying | 126 | 1.02 | (1.22) | 0.52 | 107 | 1.90 | (1.61) | 0.58 | 233 | 1.42 | (1.47) | 0.58 |
2. Frequency | 126 | 2.01 | (1.61) | 0.37 | 106 | 2.37 | (1.68) | 0.36 | 232 | 2.17 | (1.65) | 0.38 |
3. Intermittency | 126 | 0.79 | (1.34) | 0.54 | 107 | 1.41 | (1.82) | 0.56 | 233 | 1.07 | (1.61) | 0.57 |
4. Urgency | 126 | 1.28 | (1.60) | 0.35 | 107 | 1.39 | (1.67) | 0.27 | 233 | 1.33 | (1.63) | 0.30 |
5. Weak stream | 125 | 0.66 | (1.22) | 0.56 | 107 | 1.39 | (1.56) | 0.54 | 232 | 1.00 | (1.43) | 0.57 |
6. Straining | 126 | 0.46 | (1.03) | 0.40 | 107 | 0.82 | (1.39) | 0.41 | 233 | 0.63 | (1.22) | 0.42 |
7. Nocturia | 125 | 3.38 | (1.89) | 0.19 | 105 | 3.75 | (1.65) | 0.18 | 230 | 3.55 | (1.79) | 0.20 |
Total symptom score | 124 | 9.50 | (5.88) | | 104 | 13.03 | (6.81) | | 228 | 11.11 | (6.55) | |
8. Quality of life | 126 | 3.20 | (1.89) | | 106 | 3.24 | (1.80) | | 232 | 3.22 | (1.85) | |
| | Female | | | Male | | | Overall | |
| n | Floor | Ceiling | | n | Floor | Ceiling | | n | Floor | Ceiling | |
Total symptom score | 124 | 2.40% | 0.00% | | 104 | 1.00% | 0.00% | | 228 | 1.80% | 0.00% | |
8. Quality of life score | 126 | 4.80% | 3.20% | | 106 | 4.70% | 5.70% | | 232 | 4.70% | 4.30% | |
| | Female | | | Male | | | Overall | |
Cronbach’s alpha
| | 0.68 | | | 0.70 | | | 0.71 | |
| | Female | | | Male | | | Overall | |
Intraclass correlation
| n | | | n | | | n | | |
Total symptom score | 40 | 0.75 | | 37 | 0.81 | | 77 | 0.80 | |
Quality of life score | 41 | 0.58 | | 38 | 0.84 | | 79 | 0.70 | |
Table
2 shows the results of corrected item-total correlation testing for the items on urinary symptoms. Corrected item-total correlation scores were ≥0.4 for four symptoms, but did not reach the standard of 0.4 for items on frequency, urgency and nocturia.
Table
2 shows the internal consistency and test-retest reliability of the IPSS HKv2. Cronbach’s alpha coefficient was 0.7 for the seven symptom-related items in males while 0.68 in females. Test-retest reliability was assessed in 77 patients whose global rating of change scale showed no change in their health between the baseline and 2-week interviews. Overall the ICC of the IPSS HKv2 total symptom and quality of life scores exceeded 0.7, and exceeded 0.8 in males. The ICC of the IPSS HKv2 quality of life question was only 0.58 in females.
Table
3 shows the results of Pearson’s correlation testing of total symptoms and quality of life scores against the ICIQ-UI SF, IIQ-7, SF-12v2 PCS and MCS scores. The IPSS HKv2 total symptom score had a moderate correlation with ICIQ-UI SF (Pearson’s correlation coefficient of 0.44 in females and 0.50 in males). The IPSS HKv2 quality of life score had a moderate correlation with IIQ-7 but weak correlations with SF-12v2 PCS and MCS. In males, there was no significant correlation between the IPSS HKv2 quality of life score and SF-12 PCS.
Table 3
Correlations of the IPSS (HK v2) with ICIQ-UI SF, IIQ-7, SF 12 v2
Female | The IPSS total symptom score | 0.44** | | | |
| The IPSS quality of life score | | 0.50** | −0.22* | −0.18* |
Male | The IPSS total symptom score | 0.50** | | | |
| The IPSS quality of life score | | 0.40** | −0.10 | −0.28** |
Overall | The IPSS total symptom score | 0.47** | | | |
| The IPSS quality of life score | | 0.46** | −0.17* | −0.21** |
Table
4 shows the sensitivity of the IPSS HKv2 in detecting differences between subjects recruited from continence care clinics and those recruited from general outpatient clinic waiting rooms. As expected, statistically significant differences were detected between the two groups for IPSS HKv2 total symptom score (effect size 0.34) and quality of life scores (effect size 0.64). Similarly, significant differences were detected between the groups for ICIQ-UI SF (effect size 0.38) and IIQ-7 (effect size 0.30). There was no significant difference between groups for SF-12v2 PCS and MCS.
Table 4
Sensitivity in detecting differences between subjects from CC and GOPC
The IPSS total symptom score | 12.33 | (6.36) | 10.16 | (6.56) | 0.013 | 0.34 |
The IPSS quality of life score | 3.86 | (1.67) | 2.74 | (1.83) | <0.001 | 0.64 |
The ICIQ-UI SF | 8.48 | (3.99) | 6.94 | (4.01) | 0.004 | 0.38 |
The IIQ-7 | 3.77 | (4.13) | 2.59 | (3.75) | <0.024 | 0.30 |
The SF-12 PCS | 46.46 | (9.14) | 46.08 | (9.58) | 0.764 | 0.04 |
The SF-12 MCS | 52.59 | (10.88) | 53.44 | (10.26) | 0.545 | 0.08 |
Back translation of the IPSS (Hong Kong Chinese version 1) identified several questions and response options which were not equivalent to the original English instrument. First, “
have you” instead of “
how often” is used for every question in the Hong Kong version 1. When respondents first read the question, they might only focus on whether they have the particular symptom and may not expect to consider the frequency of the symptom, causing discrepancies in how they might respond. Second, the meaning of “
the rest of your life” is missed in the quality of life question in the Hong Kong Chinese version 1. Its purpose in the original instrument is to ask respondents to consider how they would feel if their urinary problems lasted until they died. Without the meaning of “
the rest of your life”, the respondent might only consider the current or short-term impact of their urinary problem. Such translational discrepancies threatens the validity of data[
20], and can affect its cross-cultural interpretability. This is the first study to support the content validity of the IPSS for both male and female patients LUTS. Both males and females could understand and correctly interpret the question items and response options. It confirmed that the questions of the IPSS are not gender specific.
In item-total correlation testing, the symptom-related items on frequency, urgency and nocturia had poor correlation in females and males. These three symptoms are predominantly storage symptoms and may suggest that these items are measuring a related but slightly different domain than the other items of the IPSS HKv2.
The IPSS HKv2 total symptom score moderately correlated with the ICIQ-UI SF which confirms that the constructs of both measures are related but not equivalent. The IPSS quality of life question had a moderate correlation with IIQ-7 but weak correlation with SF-12v2 suggesting that there is a difference in the construct of the IPSS quality of life question and that of the SF-12v2. Since SF-12v2 is a generic health-related quality of life measure, the domains of such measures might not be specific and sensitive enough to capture the impact of LUTS on health-related quality of life. Generic measures contain irrelevant domains and may miss specific concerns held by the respondents[
21]. On the contrary, the IIQ-7 is a condition-specific measure, so the domains of IIQ-7 should be more relevant to those with LUTS. Our results support the added value of condition-specific measures.
Interestingly, the IPSS HKv2 quality of life score only correlated with SF-12v2 MCS in men whilst there is a stronger correlation between IPSS HKv2 quality of life score and SF-12v2 PCS in women. This shows that the impact of LUTS on quality of life is different for males and females. The present study showed that LUTS appears to have a greater impact on mental health than physical health in men, whereas the opposite occurs in women. These findings are consistent with the validation study of the Spanish IPSS[
3]. Males with LUTS are worried and embarrassed[
22], and have concerns about sexual competence and prostate cancer[
23]. Men with LUTS might therefore have more mental health burden. Conversely, in women, overactive symptoms such as urgency and frequency appear to have a more negative impact on physical functioning than mental health[
24,
25].
The IPSS HKv2 was found to be a reliable patient reported outcome measure in both males and females. Overall, the internal consistency (Cronbach’s alpha >0.7) of the IPSS HKv2 appears to be comparable to the original English instrument[
1]. The Cronbach’s alpha was slightly below standard in females, which is still acceptable. The 2-week test-retest reliability in patients with stable health condition was acceptable and comparable to other versions of the IPSS[
1,
3,
26]. The intra-class correlation coefficients of IPSS HKv2 total symptom and quality of life scores were above 0.7 with the exception of the quality of life question in female subjects (ICC = 0.578) implying that this single-item measure is comparatively unreliable in women[
12]. Clinicians and researchers should interpret the responses to this item with caution and should not rely on the IPSS single-item quality of life question to monitor patients longitudinally or for evaluation of treatment effect.
The finding of higher IPSS HKv2 total symptom severity and poorer health-related quality of life in patients recruited from continence clinics than general outpatient clinics shows that the instrument is sensitive to differentiate patients with varying disease severity. It should be noted that the IPSS quality of life question had a larger effect size than all the other measures suggesting that this single-item measure may have high sensitivity to differentiate subgroups. Our findings also suggest that an individual’s perceived impact on health-related quality of life might be a determinant for patients to seek medical advice. From our data, it appears that the SF-12v2 may not be sensitive enough to detect differences between patients in continence clinics and general outpatient clinics. The SF-12v2 may be too generic and not sufficiently sensitive to detect subgroup differences[
27,
28]. Our findings indicate that condition-specific measures outperform generic measures for detecting differences between subgroups of patients with urinary symptoms.