| Popis: |
The objective of the present study was to propose a new method for the calculation of estimated hidden prevalence (EHP) in Hansen's disease (HD). We analyzed the records of 4142 HD patients diagnosed in the state of Rio Grande do Sul, Brazil, between 1970 and 1991. Out of these 4142 cases, 3291 patients had their grade of disability (GD) evaluated at the time of diagnosis and provided information about the time elapsed between the appearance of the symptoms and the moment when HD was identified by a physician (diagnosis delay, DD). Mean DD for the sample (in years) was 1.51 for disability grade 0, 2.14 for grade 1, 4.46 for grade 2, and 9.64 for grade 2. EHP was calculated taking into account only two strata of GD using the formula HP = [(NC-GD 0/1) x 2.0 + (CN-GD 2/3) x 5.0]/(CGE x PCP), where: NC-GD 0/1 = mean annual number of newly detected grades 0 or 1 cases; CN-GD 2/3 = mean annual number of newly detected grades 2 or 3 cases; CGE = proportion of newly detected cases with GD evaluated; PCP = proportion of the population covered by the state HD control program; 2.0 and 5.0 correspond to an approximation of the mean time in years of DD in each respective stratum of GD. Applying this model, we found an EHP of 529 cases which translates to an excess of 0.58 cases/10,000 population. We also conducted a multivariate analysis using a logistic regression model. This analysis revealed that, in addition to DD, other variables such as clinical form, age group, sex and mode of detection were independent risk factors for the presence of disabilities. We also found two significant effect modification factors: DD versus clinical form and DD versus age group. Taking these findings into consideration, a more complex model was used to calculate the EHP with 16 strata (defined by clinical form of the disease, age group, and GD from 0 to 3). An EHP of 502 cases (excess of 0.55/10,000) was obtained with this more complex model. This result differs only 5% from that of the simplified model. Therefore, we conclude that the simplified model is indicated to estimate hidden prevalence of HD in the field. |
