Article Text

Download PDFPDF

Patients lost to follow up: experience of an HIV clinic
  1. L J Haddow,
  2. S G Edwards1,
  3. K Sinka2,
  4. D E Mercey3
  1. 1Department of Genitourinary medicine, Camden Primary Care NHS Trust, London, UK
  2. 2Communicable Disease Surveillance Centre, PHLS Colindale, London, UK
  3. 3Department of Sexually Transmitted Diseases, Royal Free and University College London Medical School, London, UK
  1. Correspondence to:
    Lewis Haddow, Department of Genitourinary Medicine, 5th Floor, Mortimer Market Centre, Capper Street, off Tottenham Court Road, London WC1E 6AU, UK; lewis.haddow{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The National Strategy for Sexual Health and HIV aims to reduce the pool of undiagnosed HIV infection in the United Kingdom.1 Potential benefits of earlier diagnosis include timely initiation of highly active antiretroviral therapy (HAART), prevention of complications of HIV, screening for STIs that are known to enhance HIV infectivity,2 and psychological support. Patients may not realise these benefits if they are lost to follow up (LFU). Previous studies have found associations between frequent non-attendance (as distinct from LFU) and less severe illness,3 drug addiction,4 and patients’ health beliefs.5

We studied the case notes of all surviving patients who had enrolled in our HIV clinic within a 15 month period but had not received medical care for 12 months or more. Patients were excluded if they had been transferred to other centres or if the case notes were unavailable. For each case, one control was matched for date of first attendance. Data including demographics, virological and immunological markers, antiretroviral therapy, and psychological and social factors were collected from the notes using a standardised proforma.

Ninety four cases were found. LFU patients were younger than controls (table 1), with a trend towards more patients being born outside the United Kingdom. Cases were about half as likely to be on HAART than controls (RR 0.46, 95% CI 0.32 to 0.66). This association was true regardless of disease stage. Numbers were too small to analyse any association between LFU and poor adherence to HAART. There were no statistically significant differences in sex, ethnicity, disease stage, or surrogate markers. The number of general practitioners with whom communication was maintained was equally low in both groups.

The number of cases with a history of psychiatric illness, substance abuse, deliberate self harm, or use of counselling or psychiatric services was not significantly different from controls. Further data provided by CDSC showed that at least 29 cases (31%) had attended another clinic for follow up, without correspondence being made between centres. These included only eight of 26 (30%) cases on HAART and eight of 22 (36%) cases with CDC stage C or CD4 count under 200 cells ×106/l. (Soundex code and date of birth were used for matching records without compromising patient identity, and subsequent treatment locations were not specified.)

This study highlights that patients who are LFU include those at all stages of disease and are not necessarily those with a lack of clinical need. It is of interest that the association between not being on HAART and being LFU is independent of clinical stage. The patients who discontinued care from our centre were a diverse group in terms of illness, ethnicity, and transmission category, typical of the clinic population as a whole.

In a patient who is symptom free and not on HAART, one might argue that a year without specialist follow up is of no clinical importance. Such patients might be better managed in primary care or in a setting which focuses less on the patient’s disease state than on their wellbeing. An appropriately designed study might further elucidate reasons that lead patients to default from follow up. Interventions need to be in place to prevent loss of follow up of patients who are at high risk of disease progression or who are on HAART.

Table 1

Comparison of characteristics of patients lost to follow up (“cases”) with matched controls. Statistical analysis by Mann-Whitney test or χ2 test


Nina Fudge helped to retrieve the data. Andrew Copas provided statistical advice.

LH developed the study, retrieved and analyzed the data, and co-wrote the text; SE co-wrote the text; DM conceived the study and provided comments on the text; KS provided further data from CDSC.