Page 227 - PhD GT

at the central agency and it may be that, of the 38 individuals that it was not possible to trace at three months (see Chapter 7 section 7.3.3), some had given an inaccurate or a false address at the outset. Some attempt could be made to ascertain the stability of the accommodation when seeking consent for follow-up and routine recording of telephone numbers both for study participants and locators could be established. Family members may be the most stable locator and study participants could be more actively encouraged to give consent for them to be contacted.
In the present study, 52% of the 79 participants not seen at three month follow-up were accounted for leaving 38 individuals or 16.5% of the initial sample unaccounted for at this point (see Chapter 7 section 7.3.3). At the twelve month follow-up point, 41 individuals or 17.8% were unaccounted for (see Chapter 7 section 7.3.4) with some overlap between this number and the number not accounted for at 3 months. Thus the proportion of study participants for whom no information at all was available was relatively small. Additional resource for follow-up would have made possible additional attempts to locate people where there was reason to think they were could be found at their address at some time. If prior permission were sought with the initial consent for follow-up, further attempts could be made through locators, through general practice and police records. In Chapter 7 the success of further follow-up attempts was described: there were diminishing returns with the increasing number of attempts made. Improved quality of information collected at intake could improve the results of additional follow-up attempts.
Description of the sub-set of participants who were followed up and comparison with those who were not provides the basis for qualifying the applicability of the study findings. Differences found on baseline measures between those who were and were not seen for follow-up are described in Chapter 8. The overall conclusion from the comparisons of those seen and not seen at three months (t2) and at twelve months (t3) were that women were more likely to be followed up than men and greater proportions of the alcohol group were seen than of the heroin group. Those not seen at t2 and at t3 were significantly younger than those seen. Baseline dependence scores were significantly higher for those not seen at t2 compared with those seen and baseline social satisfaction scores suggested greater social dissatisfaction at t1 in those not seen at twelve months compared with those seen at twelve months. Thus, where significant differences in baseline data were found, the trend was towards greater severity in the participants who were not located or not available for follow-up. There may be a bias in the study findings which operates against younger males and heroin users who have a greater severity of problems and do not engage in treatment.
Incomplete responses to items in the questionnaire battery may render some of the scales unusable for analysis. Those supervising the completion of the questionnaires were instructed to
222