avertablemortality/processor.py at main · leonchlon/avertablemortality · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
import numpy as np
import pandas as pd


class Processor:
    @staticmethod
    def process_measure(data, measure, benchmark):
        """
        Process data for a specific measure.

        Args:
            data (pd.DataFrame): The aggregated data.
            measure (str): The measure to process (e.g., 'Deaths', 'DALYs').
            hic_countries (list): List of high-income countries for benchmarks.

        Returns:
            pd.DataFrame: Processed data for the specified measure.
        """
        if measure == 'Deaths':
            return Processor._process_deaths(data, benchmark)
        else:
            return Processor._process_other_measure(data, measure, benchmark)

    @staticmethod
    def _process_deaths(data, benchmark = 'global_benchmark'):
        """
        Process data for the 'Deaths' measure, including counterfactual calculations.

        Args:
            data (pd.DataFrame): The aggregated data.
            hic_countries (list): List of high-income countries for benchmarks.

        Returns:
            pd.DataFrame: Processed data for 'Deaths'.
        """
        final_data = None
        regional_le = data[['regional_benchmark','global_benchmark','Location']].drop_duplicates()
        regional_le = regional_le.set_index('Location')
        for sub_measure in ['Deaths', 'Prevalence']:
            measure_data = data[data['Measure'] == sub_measure].set_index(['Location', 'Sex', 'Age', 'Cause','region'])
            measure_data = measure_data.pivot(columns='Metric', values='Value')
            measure_data.columns = [f'{sub_measure}', f'{sub_measure} Rate']
            if final_data is None:
                final_data = measure_data
            else:
                final_data = final_data.merge(measure_data, left_index=True, right_index=True, how='inner')
        final_data = final_data.merge(regional_le,
                                      left_on=['Location'],
                                      right_index=True)

        # Calculate HIC mean counterfactual CF
        gpby = ['Sex', 'Age', 'Cause']
        ret_cols = ['Deaths', 'Counterfactual CF', 'Adjusted Deaths', 'Avertable Deaths']
        if benchmark == 'regional_benchmark':
            gpby += ['region']

        hic_mean_cf = (
            final_data.loc[final_data[benchmark]]
            .groupby(gpby)
            .apply(lambda x: np.nanmean(x['Deaths Rate'] / x['Prevalence Rate']))
            .rename("Counterfactual CF")
        )

        final_data = final_data.join(hic_mean_cf, on=gpby, how='inner')
        final_data = final_data[final_data[['Prevalence', 'Deaths']].min(axis=1) > 10]

        # Calculate adjusted deaths and avertable deaths
        final_data['Adjusted Deaths'] = final_data['Prevalence'] * final_data['Counterfactual CF']
        final_data['Avertable Deaths'] = final_data['Deaths'] - final_data['Adjusted Deaths']
        final_data['Avertable Deaths'] = np.where(final_data['Avertable Deaths'] < 0, 0, final_data['Avertable Deaths'])

        return final_data[ret_cols]

    @staticmethod
    def _process_other_measure(data, measure, benchmark):
        """
        Process data for measures other than 'Deaths'.

        Args:
            data (pd.DataFrame): The aggregated data.
            measure (str): The measure to process.
            hic_countries (list): List of high-income countries for benchmarks.

        Returns:
            pd.DataFrame: Processed data for the specified measure.
        """
        regional_le = data[['regional_benchmark','global_benchmark','Location']].drop_duplicates()
        regional_le = regional_le.set_index('Location')

        measure_data = data[data['Measure'] == measure].set_index(['Location', 'Sex', 'Age', 'Cause','region'])

        # Pivot data to get metrics as columns
        pivoted_data = measure_data.pivot(columns='Metric', values='Value')
        pivoted_data.columns = [measure, "Rate"]

        # Filter out rows with very low values
        filtered_data = pivoted_data[pivoted_data[measure] > 10]

        filtered_data = filtered_data.merge(regional_le,left_on=['Location'], right_index=True)

        # Calculate HIC mean counterfactual CF
        gpby = ['Sex', 'Age', 'Cause']
        ret_cols = [measure, f'AdjustRatio {measure}', f'Adjusted {measure}', f'Avertable {measure}']
        if benchmark == 'regional_benchmark':
            gpby += ['region']

        # Calculate HIC mean rate
        hic_mean_rate = (
            filtered_data.loc[filtered_data[benchmark]]
            .groupby(gpby)['Rate']
            .mean()
            .rename('HIC_mean_rate')
        )
        filtered_data = filtered_data.join(hic_mean_rate, on=gpby,how='inner')
        # Calculate adjustment ratio
        filtered_data[f'AdjustRatio {measure}'] = (filtered_data['HIC_mean_rate'] / filtered_data['Rate']).clip(upper=1)

        # Calculate adjusted values and avertable values
        filtered_data[f'Adjusted {measure}'] = filtered_data[measure] * filtered_data[f'AdjustRatio {measure}']
        filtered_data[f'Avertable {measure}'] = filtered_data[measure] - filtered_data[f'Adjusted {measure}']
        return filtered_data[ret_cols]