Table 3.
The association between data discrepancy (between clinics' outpatient books and government statistics in clinical malaria case reporting) and variables, using multiple regression.
| Variable | Coefficient | (95%CI§) | P-value |
| (a) Average numbers of patients (per nurse and day) | 0.05 | (0.31,0.07) | < 0.001 |
| (b) Illegible handwriting (per 100 patients) | 0.09 | (0.04,0.15) | 0.017 |
| (c) Omitted data of diagnosis(per 100 patients) | 0.04 | (-0.06,0.13) | 0.642 |
| (d) Omitted data of slide confirmation(per 100 patients) | 0.02 | (-0.03,0.07) | 0.567 |
| (e) Omitted data of treatment (per 100 patients) | 0.06 | (-0.05, 0.18) | 0.447 |
| (f) Registered nurses among all nurses and nurse aides (%) | -0.01 | (-0.02, 0.01) | 0.398 |
| (h) Use of tally sheets† | -0.38 | (-0.54, -0.22) | < 0.001 |
| (j) Shared official definition of clinical malaria by nurses | -0.07 | (-0.28, 0.13) | 0.379 |
| (l) Clinic level╙ | -0.48 | (-0.89, -0.06) | 0.019 |
| (m) Rainy season∏ | -0.05. | (-0.34, 0.24) | 0.75 |
(n = 92)
(Alphabet) is correspondent to the text.
Significant variables are in bold. §Confidence Interval (Adjusted R squared: 0.67)
(a)-(f) are continuous, (h)(j) are ordinal, and (l)(m) are categorical variables. The following variables were removed because of the violation to multicollinearity and overfitting: (g) whether registered nurse filled in reports, (i) daily counts of cases, and (k) whether weekly meetings occurred.
†Tally sheets are the sheets with a lot of blank circles divided into ten to help nurses to count cases by marking circles according to the records in outpatient books. They were distributed by the government to all clinics.
╙Two clinics, Kukum and Rove, are upper clinics called Area Health Clinics (AHC) given more human and material resources by the government, compared to other six clinics called Urban Health Clinics (UHC).
∏ From November to March