Family & friends in social networks: us & them the US....or Some Thoughts on Families and Social Networks based on UKATT data New Directions in the Study of Alcohol Group – Alcohol Spills Over – Birmingham, 37th Annual International Conference – April 2013 Aims for my presentation To provide an overview of social network composition of people entering alcohol treatment using UKATT IPA data To use UKATT data on who are the most important people for those with alcohol problems entering treatment to provide a very cautious estimate of number of significant others affected by alcohol problems To explore the potential role of friends (drinking and nondrinking) in influencing outcomes from alcohol treatment – describe previous work and some preliminary analyses Finally I will briefly mention 3 trials I am involved in testing in some way social network interventions as a form of help for people using substances UKATT Research Team (Principal Investigators) • • • • • • • • Alex Copello Christine Godfrey Nick Heather Ray Hodgson Jim Orford Duncan Raistrick Ian Russell Gillian Tober West Midlands York Newcastle South Wales Birmingham Leeds York/Bangor Leeds Seventy five percent of the UKATT PI Network Important People and Activities Inventory (IPA) Clifford and Longabaugh, 1991 IPA asks the participant to nominate up to 10 people with whom the client has spent most time in the last 3 months. Then it also asks to indicate the 4 most important from those nominated. IPA then collects detailed information about each of these people e.g. frequency of contact, drinking frequency and quantity, how much the person is liked, how encouraging and supportive of not drinking and coming to treatment Participants entering the UKATT trial were asked about the people they had spent most time with over the previous three months (aged at least 12 years old). They named these ‘important’ people and described each of them using demographic and drinking-related characteristics. Overall 4677 important people were named. All 740 participants could name at least 1 important person; the highest number named was 12. Participants could most frequently name 10 people, and the mean number named was 6.5 (standard deviation 3.6). This graph shows numbers of important people named by participants: 120 Participants 100 80 60 40 20 0 1 2 3 4 5 6 7 8 9 10 11 12 Important people Participants then decided which of these people were the four most important and listed them in order. The four most important people: how were they related to focal drinkers? This graph shows how the participants were related to their four most important people. Percentages of named relationships are shown for the 1st, 2nd, 3rd and 4th most important people to the focal drinker: 70 partner/ex-partner 60 50 40 fam ily eg. child, parent, s ibling 30 friend 20 other 10 0 1s t 2nd 3rd 4th The majority of focal drinkers named their partner as their most important person, although family members were also a frequent choice. Very few focal drinkers named their partner as least important person of the four. Members of close family were frequently named as second or third most important, and friends were commonly named as third or fourth. Who were the most important family members? Relationships of important people to focal drinkers could be further broken down into relationships of family members. This graph shows percentages of type of family member (from overall relationship) who were named as each important person: 25 20 children 15 parent s ibling 10 other/family 5 0 1s t 2nd 3rd 4th Parents were most frequently named as first and second most important, children as second or third most important, and siblings as third most important. Other family members were more likely to be named as the fourth than as the most important person. Were important people male or female? The most important person named by focal drinkers was more commonly female; two-thirds of people named were female (66.58%) and a third were male (33.42%). The second, third, and fourth most important people named by focal drinkers were equally likely to be either male or female. Graph showing percentage of males and females that were named as each important person (1st, 2nd, 3rd, or 4th most important): 70 60 50 40 male 30 female 20 10 0 1st 2nd 3rd 4th How often were they seen by the focal drinkers? Participants saw their most important person frequently, with over half (58.92%) seeing them every day. Many also saw their second most important person every day (34.51%). The third most important person was most frequently seen 3-6 times a week (32.71%). Fourth most important people were often seen 1-2 times a week (34.52%) or 3-6 times a week (32.45%). It was rare for any important people not to have been seen in the past 3 months (under 1% for each person). 70 60 50 40 30 20 10 0 1st 2nd not at all in past 3 months once in past 3 months less than monthly about once a month every other week 1-2 times a week 3-6 times a week every day 3rd 4th What was their drinking status? All four important people were most often considered ‘light drinkers’. They were also more often moderate drinkers than abstainers, although abstainers made up 16.8% of four most important people. The most important people were the most frequent abstainers (20.84%), and the least frequent heavy drinkers (10.83%). This graph shows the percentages of important people (1st, 2nd, 3rd and 4th) who fell into each drinking category: 45 40 35 30 25 20 15 10 5 0 1st 2nd 3rd 4th recovering abstainer alcoholic light drinker moderate drinker heavy drinker don't know How often did they drink alcohol? Important people often drank regularly. All important people most commonly drank alcohol ‘1-2 times a week’, although the most important people less frequently drank ‘1-2 times a week’ (24.22%) than the others. The most important people were also described as having ‘not drunk alcohol in the past 3 months’ more frequently than were the other three people. Second and third most important people were the most frequent daily drinkers (17.46; 17.21%). This graph shows percentages of important people (1st, 2nd, 3rd, 4th) in each drinking category: 30 25 1st 20 2nd 15 3rd 10 4th 5 0 not in past 3 months once in past 3 months about once a month about every other w eek 1-2 times 3-6 times per per w eek w eek daily don't know How much were they liked by the focal drinkers? 60 50 1s t 40 2nd 30 3rd 20 4th 10 0 totally liked very m uch quite a bit m ixed feelings dis liked dis liked a lot totally dis liked How important were they to the focal drinkers? 60 50 1s t 40 2nd 30 3rd 20 4th 10 0 extrem ely im portant very im portant im portant s om ewhat im portant not very im portant not at all im portant How did they react to the focal clients’ drinking? 60 50 1st 40 2nd 30 3rd 20 4th 10 0 left or made you leave didn't accept neutral accepted encouraged don't know How did they react to them not drinking? 70 60 50 1st 40 2nd 30 3rd 20 4th 10 0 left or made you leave didn't accept neutral accepted encouraged don't know How did they feel about the focal drinker getting alcohol treatment? Important people rarely opposed treatment. The most important person was particularly supportive (78.89% supported treatment), and even fourth most important people were predominantly supportive (53.18%). Just over 10% of all important people felt neutral towards treatment. The third and fourth most important people were more likely than the first/second not to know about treatment. This graph shows the percentages of important people (1st, 2nd, 3rd and 4th who fell into each category: 90 80 70 60 50 40 30 20 10 0 1st 2nd 3rd 4th opposes treatment mixed neutral supports but supports prefers treament in alternative this program doesn't know So, overall.... • Participants recruited to UKATT had people in their social networks • Based on participants’ reports, most had people in their networks that did not have alcohol problems • Most had people in their networks that were supportive of the focal client’s treatment Day, Copello, Chohan et al., 2012. European Addiction Research Opiate Substitution Treatment in the UK 118 participants identified a total of 820 network members Mean network size was 6.9. Of this group, 47 378 97 189 16 51 42 (6%) were sexual partners (46%) immediate family members (12%) extended family members (23%) friends (2%) colleagues from work (6%) treatment professionals or members of self-help groups (5%) others So, two thirds of the social network was made up of family members Social Behaviour and Network Therapy SBNT involves bringing members of the focal drinker’s social network into their treatment. The UKATT data indicated that all 740 participants had at least one person who was important to them. SBNT requires that these people be: 1. free from alcohol problems 2. in regular contact with the focal drinker Further analysis of the data suggested there were 716 focal drinkers with an important person who drank no more than moderately, and who they saw at least weekly. This means that approximately 96.7% of all focal drinkers entering UKATT may have had a potential network member. Orford (Potential Supportive Network Members) • Detailed analysis of 50 cases of UKATT • 39 (78%) had at least one IP judged Definitely Suitable (14 had 1; 13 had 2; 11 had 3 and 1 had 4) • 9 (18%) others with al least one IP judged Probably Suitable • 2 (4%) others with one IP judged Possibly Suitable Important People Can we have a go at estimating numbers of people affected by those entering UK treatment and who they are? The most important people likely to be affected For 740 clients entering treatment for alcohol problems: Of the four most important people nominated: Included those people who had daily contact with the focal drinker Excluded those people who were ‘heavy drinkers’ themselves or the drinking status was ‘unknown’ 1st Imp.Person 397 (45 exc.) 2nd Imp.Person - 205 (40 exc.) 3rd Imp.Person 140 (21 exc.) 4th Imp.Person – 84 (21 exc.) Total 826 (127 exc.) The 4 most important people who have daily contact with the focal drinker Chart Title 500 450 400 350 300 heavy drinkers 250 Non-heavy drinkers 200 150 100 50 0 1st IP 2nd IP 3rd IP 4th IP This suggests that a very cautious estimate of people significantly affected • For every person walking through the door of an alcohol treatment agency on average slightly more than one other person can be significantly affected • For 740 focal drinkers 826 significant others are affected • Of the 826: 686 (83.05%) are family members and 140 (16.9%) friends Categories of family members who were not heavy drinkers and had daily contact with the focal drinker (N = 686) 300 270 250 200 Partners 156 Children >12 146 150 Parents Siblings Other family 100 58 56 Siblings Other family 50 0 Partners Children >12 Parents Those with whom the focal client had daily contact and were heavy drinkers Family Members Friends Adult Family Members Affected. Illegal Drugs - UKDPC work. What is the extent of the problem? • Drug treatment population General population 50,373 partners 55,012 parents 35,208 ‘other’ 573,671 partners 610,970 parents 259,133 ‘other’ • • • Total = 140,593 Total =1,443,774 Key findings from UK DPC study about adult family members of drug misusers. What about alcohol misuse? Up to 1 million children are affected by parental drug misuse & up to 3.5 million by parental alcohol misuse (Manning et al., 2009). It is estimated that the impact of drug misuse on the family costs the UK £1.8 billion but also brings a resource saving to the NHS of £747 million through the care provided. Estimates of adult family members affected in alcohol treatment population UKDPC drugs estimate – using treatment figures -2009 50,373 partners (35.8%) 55,012 parents (39.1%) 35,208 ‘other’ (22.04) _____________________________________________ Total = 140,593 UKATT alcohol estimate – using treatment figures – 2011-12 46,200 partners (43.8%) 20,818 parents (19.7%) 22,249 sons-daughters over 12 (19.2%) 16,256 ‘other’ (15.4%) ______________________________________________ Total = 105,523 Overall proportion of family-focused practice – increases after targeted program of work within an organisation (from Claire Hampson’s work) % family-focused practice * *^ 34 26 19 Baseline (117/611) Training & Supervision period (108/410) Post-intervention (257/765) *Significantly greater than baseline. ^Significantly greater than training period. ...and finally, the issue of social support.... Boca Juniors stadium - Argentina The addition of just one alcohol abstinent person to the social network may increase the likelihood of abstinence by 27% (Litt et al., 2007). ‘Getting by (or getting high) with a little help from my friends’....(Mohr et al., 2001; Journal of Studies on Alcohol) To evaluate friendships: • Boyfriends, girlfriends, spouses and family members were excluded. • Friends were then categorised according to drinking status (moderate or heavy as ‘drinking friends’ – other categories as ‘non-drinking friends’) • Proportion of drinking and non-drinking friends in the network – number of friends divided by total network members • This was done for the total number of friends and for the friends amongst the four most important people • Finally, variables for ratings of liking and importance of drinking and non-drinking friends were created MATCH Drinking Friends Non-drinking Friends Baseline 9-month Baseline 9-month % of friends in the total network 24.63 17.98 23.03 33.37 Years known friends 9.28 10.24 9.08 8.61 Amount of contact 5.40 5.20 5.48 5.32 Of all people in the network UKATT Drinking Friends Of all people in the network % of friends in the total network Years known friends Amount of contact Non-drinking Friends Baseline 12-month Baseline 12-month 23.1 22.4 17.0 20.7 10.5 11.1 8.8 9.7 5.39 5.19 5.37 5.14 Amount of contact: 0 = not at all, 1 = once in past 6 months, 2 = less than monthly, 3 = once a month, 4 = every week, 5 = 1-2 times/week, 7 = daily; MATCH Drinking Friends Of the 4 most important people % of friends in 4 most important people Liking of friends Importance of friends Non-drinking Friends 19.00 (23.53) 14.81 (23.21) 19.85 (23.51) 21.47 (27.05) 5.42 (1.10) 5.77 (1.05) 5.74 (1.01) 5.85 (1.05) 4.00 (1.10) 4.54 (1.14) 4.41 (1.08) 4.70 (1.12) UKATT Drinking Friends Of the 4 most important people % of friends in 4 most important people Liking of friends Importance of friends Non-drinking Friends Baseline 12-month Baseline 12-month 17.5 15.6 13.1 14.3 5.61 5.91 5.85 5.97 4.04 4.88 4.50 4.79 Liking: 1 = totally dislike, to 7 = totally like. Importance: 1 = not at all, to 6 = extremely. Mohr et al, 2001 –Summary of findings • Those participants who continued to have a higher proportion of drinking friends and higher proportion of important drinking friends at follow-up drank more drinks per drinking day (DDD) than those who reduced the proportion of drinking friends. • Those participants with a higher proportion of nondrinking friends at follow-up showed a reduced number of drinks per drinking day and that a higher rated importance of non-drinking friends from baseline to follow-up was associated with increased proportion of days abstinent at follow-up. Are friends important for drinking outcomes? With Dr Ed Day and Dr Clare Mackie (IoP) As a first step we have done a replication of the Mohr et el. analyses. Using the whole sample (N = 735) we examined whether 12 month friendships can mediate the relationship between baseline friendships and follow-up drinking by testing 4 models. MPLUS was used to test the models and to deal with missing data with a method that involves estimating the model with maximum likelihood (assuming that the data is missing at random) which uses all the available data in the model to provide the covariance/regression estimates, without imputing the data So far.... • Model 1: Proportion of drinking friends as a mediator with DDD and then PDA as outcome variables • Model 2: Proportion of non drinking friends as a mediator with DDD and then PDA as outcome variables • Model 3: Proportion of important drinking friends as a mediator with DDD and then PDA as outcome variables • Model 4: Proportion of important non-drinking friends as a mediator with DDD and then PDA as outcome variables Control variables * Baseline drinking variables: DDD and PDA * Treatment condition (MET or SBNT) * Demographics: age, gender, marital status, education, ethnicity Proportion of drinking friends – follow-up Proportion of drinking friends - Baseline Percentage Days Abstinence - PDA What we found so far.... Of the 4 models tested the only significant mediation models involved proportion of drinking friends and proportion of important drinking friends on PDA at 12 months Current Research Studies with family and social network components in Birmingham and beyond.... • NIHR - HTA 11/60/01 Adaptation and feasibility study of a family and social network intervention for young people misusing alcohol and drugs. £528,563. 30 months from 06/13. • NIHR - A pilot study of a social network intervention for heroin users in opiate substitution treatment. NHS Research for Patient Benefit. £250,000. 3 years from 01/12. • NIHR - A study to assess the Feasibility and impact of a motivational intervention on problem drug and alcohol use in adult mental health in-patient units. NHS Research for Patient Benefit. £234,269. 2 years from 06/12. Thank you for listening… ‘Despite the recognition of carers’ needs and the growth of carer organisations, however, there is a rather limited evidence base assessing the impact on carers and families of people who misuse drugs and on interventions intended to support them, and even less attention given to the needs of families and carers in their own right.’ Nice 2008.