Development and validation of a new treatment outcome score for men with LUTS.

AIMS

To develop, validate, and test elements of a new outcome score for men with lower urinary tract symptoms (LUTS).

METHODS

Elements of well-established questions from the International Prostate Symptom Score (IPSS) assessment were combined with both established noninvasive objective determinants of voiding function including maximum uroflow (Qmax), postvoid residual urine volume (PVR), total number voids, maximum voided volume (MVV) as well as a subjective patient assessment parameter, to create a new LUTS treatment outcome instrument which we have termed the "LUTS outcome score" (LOS). The LOS is comprised of eight items; each assigned a score of 0, 1, or 2. Thus, the score ranges from 0 (best) to 16 (worst). Patients were divided into surgical/non- surgical (pharmacologic or behavior modification) treatment groups. Content validity and cutoff values for cure/improve/fail were established by an expert panel. Criterion validity was established by comparison to the IPSS. Internal reliability analysis was performed to obtain information about the relationships between individual items in the scale. Individual LOS items were correlated with the subjective outcome score. We also calculated the correlations between the LOS, IPSS, subjective post-treatment response, and age. Internal consistency, based on the average inter-item correlation was calculated using Cronbach's alpha statistic. Associations between continuous variables were examined by calculating the Pearson correlation coefficient, and between ordinal variables, using Spearman's rho. Test-retest analysis was recorded to determine reproducibility of the patient subjective outcome score. Seventy-seven men who underwent treatment for LUTS for at least 6 months participated in the study. All completed a detailed pretreatment and post-treatment clinical evaluation consisting of history/physical, questionnaire, voiding diary, urinalysis, Q, PVR, and videourodynamic study. Subjective responses of cured/improved/failed status following treatment were assessed by independent investigators.

RESULTS

A high level of internal consistency was observed among the LOS symptom questions, Cronbach's alpha = 0.81 for the total cohort. The LOS correlated directly and significantly with the patient's subjective post-treatment response (r = 0.75, P < 0.001), age (r = 0.27, P = 0.02), and with the total IPSS (r = 0.62, P < 0.001). The IPSS also correlated directly and significantly with the patient's subjective post-treatment response (r = 0.48, P < 0.001) but did not correlate with the patient's age (r = 0.10, P = 0.41). When the subjective parameter for post-treatment response was subtracted from the LOS, the resulting correlation (r = 0.52, P < 0.001) approximated that obtained using the IPSS.

CONCLUSIONS

The LOS is valid and internally consistent. Unlike the IPSS, it combines subjective, semisubjective, and objective parameters. Because of the objective components, it diminishes the possibility of overestimating cure and underestimating improvement relying only on patient's symptoms. Further, we believe it will prove to be useful for post hoc analysis in patients who have not had pre-treatment assessment.