Case-management by the GP of domestic violence an example of results from a sentinel network of general practitioners Nathalie Bossuyt Sentinel network of general practitioners
Outline Introduction Research questions Methods Results Conclusions
Introduction: WHO classification violence selfdirected interpersonal collective selfabuse family / partner community social suicidal behaviour adult victim acquaintance political elderly victim stranger economic child victim
Introduction: WHO classification violence selfdirected interpersonal collective selfabuse family / partner community social suicidal behaviour adult victim acquaintance political elderly victim stranger economic child victim
Outline Introduction Research questions Methods Results Conclusions
Research Questions Characteristics of cases of domestic violence seen by the GP Case-management of domestic violence by the GP Practice variation in case-management
Outline Introduction Research questions Methods Results Conclusions
The sentinel network of GPs a network of practices or community based physicians who monitor one or more specific health problems on a regular basis representative for all general practitioners in the country (age, sex, geographical distribution) operational for 30 years coordinated at IPH voluntary participation regular participation small financial compensation + (individual) feedback +/- 8 health problems registered / year weekly registration / paper forms anonymity of patients
Data collection & study population Data collection: 01/01/2002 31/12/2004 Inclusion: all cases of interpersonal violence Topics: Socio-demographics victim, perpetrator, p household GP: Socio-demographics + practice organisation Nature of violence (physical, psychological, sexual) Place of violence Reporter of complaint Provision of care by GP: interventions / referrals (long term) consequences for victim
Paper form
Statistical analysis For each (binary) intervention or referral: Multilevel logistic regression analysis Random effects model Why? Data on patients are clustered by GPs Multilevel analysis takes into account clustering valid estimates of regression parameters correct confidence intervals & inferences Random effects model: estimates variation between GPs
Dependent variables: interventions / referrals Interventions Next appointment Prescription psychotropic medication Certificate of injury Certificate of incapacity of work Hospitalisation No immediate interventions Referrals Child abuse centre Guidance centre Shelter / safe house Social services Police / juridical services Psychologist / psychiatrist Other GP / specialist Any referral (incl. hospitalisation)
Independent variables: individual & GP characteristics Individual characteristics Socio-demographics victim Socio-demographics perpetrator Socio-demographics household Nature of violence Place Reporter of complaint Other interventions/referrals GP characteristics + region Region Age & sex GP Group/solo practice Size of practice Trainee Organisation patient s t record
Outcome measures for each intervention Frequency of occurrence: % calculated on basis of parameters of random effects model Variation between GPs: Median Odds Ratio: median of all odds ratios resulting from every comparison of two persons with the same covariates but from randomly chosen different GPs always > 1 the higher the MOR, the larger the variability between clusters
Outline Introduction Research questions Methods Results Conclusions
Size of the network Number of GP practices 2002 150 2003 163 2004 181 Population covered 2002 15% 1.5% 2003 1.6% 2004 1.8% Registered cases of interpersonal violence N = 1745 Registered cases of domestic violence N = 963 Incidence of GP consult. for intrafamilial violence: 2 / 1000
Number of registered cases violence interpersonal (N = 1745) collective selfdirected selfabuse family / community partner (N = 733) (N = 963) social suicidal behaviour adult victim (N = 776) acquaintance (N = 403) political elderly victim (N = 86) stranger (N = 330) economic child victim (N = 73)
Demographic characteristics Male victims Female p-value victims N(%) N (%) Age group victim 17y 34 (19) 47 (6) 18-64y 117(65) 648 (86) 65y 29(16) 57 (8) < 0.0001 Sex perpetrator t man 95(66) 609 (95) woman 49 (34) 34 (5) < 0.0001 Self-reported complaint yes 146 (78) 693 (90) < 0.0001 history of domestic violence Previous episodes 89 (48) 492 (64) First episode 96 (52) 274 (36) < 0.0001
Characteristics of the violence Male Female p-value victims victims N (%) N (%) Nature of the violence Physical violence 150 (80) 586 (76) 0.25 Psychological violence 134 (71) 636 (82) 0.0006 Sexual violence 8 (4) 53 (7) 0.24 Physical injuries No injury 20 (11) 67 (9) 0.41 Open wound 41 (22) 86 (11) 0.0001 Haematoma 80 (43) 387 (51) 0.05 Fracture 7 (4) 26 (3) 0.82 Burn wound 3 (2) 5 (1) 0.20 Contusion 38 (20) 196 (26) 013 0.13 Concussion 6 (3) 15 (2) 0.28 Psychological violence Emotional pressure 68 (37) 398 (53) 0.0001 Verbal violence 104 (57) 480 (64) 0.08 Extorsion 12 (7) 101 (13) 001 0.01 Neglect 7 (4) 34 (5) 0.84
Circumstances Male Female p-value victims victims N(%) N (%) Substance abuse acute abuse alcohol victim 13 (7) 21 (3) 0.01 chronic abuse alcohol victim 26 (14) 61 (8) 0.01 Place of the violence home 141 (75) 686 (89) < 0.0001 Other house 15 (8) 43 (6) 0.21 institution 0 (0) 6 (8) 0.60 Public place 27 (14) 47 (6) 0.0001 work 4 (2) 8 (1) 0.27 other 1 (1) 5 (1) 1.00 Long-term consequences Mental problems 53 (28) 321 (42) 0.0007 absence from work or school 19 (10) 91 (12) 0.52
Case-management of domestic in adults by the GP Interventions Frequency of Variation occurrence between GPs (%) (MOR) Next appointment 37 34 1.8 4.1 Prescription psychot. med. 11 14 3.2 2.1 Certificate t of injury 42 60 12 1.2 29 2.9 Certificate of incapacity of work 11 14 4.7 1.7 Hospitalisation ti 4 3 / 19 1.9
Case-management of domestic in adults by the GP referrals Frequency of Variation occurrence between GPs (%) (MOR) Psychosocial services 11 13 / 2.4 Police / juridical services 5 23 13.3 2.3 Other health professional / 8 12 3.9 2.7 medical service
Conclusions GP is frequently consulted for interpersonal violence Considerable practice variation for most interventions / referrals
Questions?