File size: 11,029 Bytes
06881dd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
#!/usr/bin/env python3
"""
Literature-Informed Obstetric Fistula Dataset
==============================================

Generates realistic synthetic records of obstetric fistula patients
in sub-Saharan Africa, including risk factors, fistula characteristics,
social consequences, surgical repair, and outcomes.

References (web-searched):
-----------
[1]  UNFPA 2024. Obstetric fistula caused by prolonged
     obstructed labour without timely EmONC. SSA has 50%
     of global burden. 50K-100K new cases/year.
[2]  UNFPA WCARO 2024. West/Central Africa highest rates.
     Social isolation, divorce, depression common.
[3]  PMC 2024. Fistula repair failure in SSA. Repair
     success 85-90% at specialized centres.
[4]  PubMed 1999. Early marriage, young adolescents,
     premature pregnancy → VVF. Sociomedical risk factors.
[5]  PubMed 2018. Kitovu Hospital Uganda, highest fistula
     rates globally. Incidence and causative factors.
[6]  PubMed 2024. Systematic review repair outcomes.
     Proportions of successful repair in LMICs.
"""

import numpy as np
import pandas as pd
import argparse
import os

SCENARIOS = {
    'specialized_fistula_centre': {
        'description': 'Dedicated fistula repair centre with '
                       'trained surgeon, physiotherapy, social '
                       'support (e.g., Addis Ababa Fistula Hospital, '
                       'Kitovu Uganda)',
        'repair_available': True,
        'specialist_surgeon': True,
        'physiotherapy': True,
        'social_support': True,
        'repair_success': 0.90,
    },
    'district_hospital': {
        'description': 'District hospital with visiting fistula '
                       'surgeon, basic repair capability, limited '
                       'follow-up (e.g., district hospitals DRC, '
                       'Malawi, Tanzania)',
        'repair_available': True,
        'specialist_surgeon': False,
        'physiotherapy': False,
        'social_support': False,
        'repair_success': 0.70,
    },
    'no_surgical_access': {
        'description': 'Rural community with no surgical access, '
                       'traditional remedies, referral barriers '
                       '(e.g., rural Niger, Chad, South Sudan)',
        'repair_available': False,
        'specialist_surgeon': False,
        'physiotherapy': False,
        'social_support': False,
        'repair_success': 0.0,
    },
}


def generate_dataset(n=10000, seed=42, scenario='district_hospital'):
    rng = np.random.default_rng(seed)
    sc = SCENARIOS[scenario]

    records = []

    for idx in range(n):
        rec = {'id': idx + 1}

        # ── 1. Demographics ──
        rec['age_at_presentation'] = max(14, min(65, int(rng.normal(28, 8))))
        rec['age_at_fistula_onset'] = max(13, min(rec['age_at_presentation'],
            int(rng.normal(20, 5))))
        rec['years_living_with_fistula'] = max(0,
            rec['age_at_presentation'] - rec['age_at_fistula_onset'])
        rec['age_at_marriage'] = max(10, min(30, int(rng.normal(16, 3))))
        rec['child_marriage'] = 1 if rec['age_at_marriage'] < 18 else 0
        rec['age_at_first_pregnancy'] = max(rec['age_at_marriage'],
            min(35, int(rng.normal(17, 3))))
        rec['education'] = rng.choice(
            ['none', 'primary', 'secondary', 'tertiary'],
            p=[0.45, 0.35, 0.15, 0.05])
        rec['rural'] = 1 if rng.random() < 0.80 else 0
        rec['parity'] = max(0, min(12, int(rng.exponential(2.5))))
        rec['height_cm'] = max(140, min(180, int(rng.normal(157, 6))))
        rec['short_stature'] = 1 if rec['height_cm'] < 150 else 0
        rec['bmi'] = round(max(14, min(35, rng.normal(20, 3))), 1)

        # ── 2. Obstetric history ──
        rec['labour_duration_hours'] = max(6, min(120, int(rng.exponential(24) + 12)))
        rec['prolonged_labour'] = 1 if rec['labour_duration_hours'] > 24 else 0
        rec['obstructed_labour'] = 1 if rec['labour_duration_hours'] > 18 and rng.random() < 0.85 else 0
        rec['caesarean_performed'] = 0
        if rec['obstructed_labour']:
            rec['caesarean_performed'] = 1 if rng.random() < 0.15 else 0
        rec['skilled_birth_attendant'] = 1 if rng.random() < 0.20 else 0
        rec['place_of_delivery'] = rng.choice(
            ['home', 'health_centre', 'hospital'],
            p=[0.60, 0.25, 0.15])
        rec['stillbirth'] = 1 if rng.random() < 0.70 else 0
        rec['neonatal_death'] = 0
        if not rec['stillbirth']:
            rec['neonatal_death'] = 1 if rng.random() < 0.15 else 0

        # ── 3. Fistula characteristics ──
        rec['fistula_type'] = rng.choice(
            ['vesicovaginal', 'rectovaginal', 'combined'],
            p=[0.80, 0.10, 0.10])
        rec['fistula_size'] = rng.choice(
            ['small', 'medium', 'large', 'extensive'],
            p=[0.20, 0.35, 0.30, 0.15])
        rec['urethral_involvement'] = 1 if rng.random() < 0.25 else 0
        rec['circumferential_defect'] = 1 if rec['fistula_size'] == 'extensive' and rng.random() < 0.40 else 0
        rec['scarring_severity'] = rng.choice(
            ['mild', 'moderate', 'severe'],
            p=[0.25, 0.40, 0.35])
        rec['vaginal_stenosis'] = 1 if rec['scarring_severity'] == 'severe' and rng.random() < 0.40 else 0
        rec['foot_drop'] = 1 if rec['prolonged_labour'] and rng.random() < 0.10 else 0

        # ── 4. Social consequences [1][2] ──
        rec['divorced_separated'] = 1 if rng.random() < 0.55 else 0
        rec['social_isolation'] = 1 if rng.random() < 0.65 else 0
        rec['depression'] = 1 if rng.random() < 0.60 else 0
        rec['anxiety'] = 1 if rng.random() < 0.45 else 0
        rec['economic_impact'] = rng.choice(
            ['none', 'mild', 'severe'],
            p=[0.10, 0.30, 0.60])
        rec['unable_to_work'] = 1 if rng.random() < 0.50 else 0
        rec['skin_excoriation'] = 1 if rng.random() < 0.70 else 0
        rec['recurrent_uti'] = 1 if rng.random() < 0.50 else 0
        rec['malodour'] = 1 if rng.random() < 0.80 else 0

        # ── 5. Care seeking ──
        rec['delay_to_presentation_years'] = rec['years_living_with_fistula']
        rec['previous_repair_attempts'] = max(0, min(5, int(rng.exponential(0.5))))
        rec['traditional_remedy_tried'] = 1 if rng.random() < 0.35 else 0
        rec['barrier_to_care'] = rng.choice(
            ['cost', 'distance', 'awareness', 'stigma', 'no_service', 'none'],
            p=[0.20, 0.20, 0.15, 0.15, 0.20, 0.10])
        rec['referred_by'] = rng.choice(
            ['self', 'ngo', 'health_worker', 'community', 'media'],
            p=[0.30, 0.25, 0.20, 0.15, 0.10])

        # ── 6. Surgical repair [3][6] ──
        rec['repair_performed'] = 0
        if sc['repair_available']:
            rec['repair_performed'] = 1 if rng.random() < 0.85 else 0

        rec['repair_technique'] = 'none'
        if rec['repair_performed']:
            rec['repair_technique'] = rng.choice(
                ['transvaginal', 'transabdominal', 'combined'],
                p=[0.75, 0.15, 0.10])

        rec['graft_used'] = 0
        if rec['repair_performed'] and rec['fistula_size'] in ('large', 'extensive'):
            rec['graft_used'] = 1 if rng.random() < 0.20 else 0

        rec['catheter_days'] = 0
        if rec['repair_performed']:
            rec['catheter_days'] = max(7, min(28, int(rng.normal(14, 3))))

        rec['repair_successful'] = 0
        if rec['repair_performed']:
            success_prob = sc['repair_success']
            if rec['fistula_size'] == 'extensive':
                success_prob *= 0.60
            elif rec['fistula_size'] == 'large':
                success_prob *= 0.80
            if rec['previous_repair_attempts'] > 0:
                success_prob *= 0.85
            if rec['circumferential_defect']:
                success_prob *= 0.50
            rec['repair_successful'] = 1 if rng.random() < success_prob else 0

        rec['residual_incontinence'] = 0
        if rec['repair_successful']:
            rec['residual_incontinence'] = 1 if rng.random() < 0.15 else 0

        rec['complication_post_repair'] = 0
        if rec['repair_performed']:
            rec['complication_post_repair'] = 1 if rng.random() < 0.10 else 0

        # ── 7. Rehabilitation ──
        rec['physiotherapy_received'] = 0
        if rec['repair_performed'] and sc['physiotherapy']:
            rec['physiotherapy_received'] = 1 if rng.random() < 0.70 else 0

        rec['social_reintegration_support'] = 0
        if rec['repair_performed'] and sc['social_support']:
            rec['social_reintegration_support'] = 1 if rng.random() < 0.60 else 0

        rec['livelihood_support'] = 0
        if rec['social_reintegration_support']:
            rec['livelihood_support'] = 1 if rng.random() < 0.40 else 0

        rec['continence_at_3_months'] = 0
        if rec['repair_successful'] and not rec['residual_incontinence']:
            rec['continence_at_3_months'] = 1 if rng.random() < 0.90 else 0

        rec['returned_to_community'] = 0
        if rec['repair_successful']:
            rec['returned_to_community'] = 1 if rng.random() < 0.75 else 0

        records.append(rec)

    df = pd.DataFrame(records)

    print(f"\n{'='*65}")
    print(f"Obstetric Fistula — {scenario} (n={n}, seed={seed})")
    print(f"{'='*65}")
    print(f"\n  Repair performed: {df['repair_performed'].mean()*100:.1f}%")
    print(f"  Repair success: {df[df['repair_performed']==1]['repair_successful'].mean()*100:.1f}%" if df['repair_performed'].sum() > 0 else "  No repairs")
    print(f"  Stillbirth: {df['stillbirth'].mean()*100:.1f}%")
    print(f"  Divorced: {df['divorced_separated'].mean()*100:.1f}%")
    print(f"  Depression: {df['depression'].mean()*100:.1f}%")
    print(f"  Mean years with fistula: {df['years_living_with_fistula'].mean():.1f}")

    return df


if __name__ == '__main__':
    parser = argparse.ArgumentParser(
        description='Generate obstetric fistula dataset')
    parser.add_argument('--scenario', type=str, default='district_hospital',
                        choices=list(SCENARIOS.keys()))
    parser.add_argument('--n', type=int, default=10000)
    parser.add_argument('--seed', type=int, default=42)
    parser.add_argument('--output', type=str, default=None)
    parser.add_argument('--all-scenarios', action='store_true')
    args = parser.parse_args()

    os.makedirs('data', exist_ok=True)

    if args.all_scenarios:
        for sc_name in SCENARIOS:
            df = generate_dataset(n=args.n, seed=args.seed, scenario=sc_name)
            out = os.path.join('data', f'fistula_{sc_name}.csv')
            df.to_csv(out, index=False)
            print(f"  -> Saved to {out}\n")
    else:
        df = generate_dataset(n=args.n, seed=args.seed, scenario=args.scenario)
        out = args.output or os.path.join('data', f'fistula_{args.scenario}.csv')
        df.to_csv(out, index=False)
        print(f"  -> Saved to {out}")