Richard Coker describes how a system, with a substantial coercive component even for non-infectious patients, evolved in New York based on a perception of risk which was perhaps fuelled by media hype.1 The reasons why such a system came about can, however, be appreciated from the state of tuberculosis control—or perhaps non-control—in New York in the early 1990s.

Due to a series of cuts in health funding, routine drug sensitivity testing had been stopped, support systems were slashed, and in some areas only about 10% of patients completed treatment.2 By 1992 33% of isolates were drug resistant, including 26% to isoniazid, and the rate of multidrug-resistant tuberculosis (MDRTB) was 19%.3 The expenditure in New York alone of $750 million (£500 million) with an extensive directly observed therapy (DOT) programme had reduced the MDRTB rate to 13% in 1994.4

In England and Wales tuberculosis notifications fell progressively until 1987, with a rise between 1987 and 1992 of some 20%5 to around 6000 cases a year. Drug resistance levels had remained low between 1981 and 19926 with a stable MDRTB rate of 0.6%. There has, however, been a rise in the MDRTB rate since 1993 up to 1.6%,7 with HIV positivity, ethnic minority groups, prior treatment, and residence in Greater London all being significant associations. Tuberculosis in the United Kingdom, as in many developed countries, is increasingly a disease which is localised to certain areas and population groups.8 ,9 The problems of tuberculosis control are largely limited to such high prevalence areas which make up some 20% of districts, with Greater London having the greatest number of such districts.9

The key elements of tuberculosis control in order of importance are (1) detection and treatment of cases, particularly those with sputum smear positive disease; (2) case holding which could be defined as maintaining treatment to completion; and (3) preventive measures such as chemoprophylaxis and BCG vaccination. There also needs to be adequate staffing levels of doctors and, in particular, of tuberculosis nurses/health visitors to deliver a service with those elements.10

The philosophical or ethical dilemma that Dr Coker raises is where the “balance point” between the libertarian and coercive strategies in tuberculosis management lies or, alternatively, where the rights of society in general outweigh the rights of an individual or vice versa. This varies according to the society and situation, and with the public perception of risk rather than the actual risk. In England and Wales currently, as a last resort, sections 37 and 38 of the Public Health Act allow for compulsory detention of a person with infectious tuberculosis of the respiratory tract. Compulsory treatment is not allowed so that compulsory admission is only sought in extreme circumstances to safeguard the public health. When such compulsory admission is sought, there are also the practical problems of maintaining such detention and of determining when “infectivity” ceases. Legally compulsory detention is only allowed for “infectious” tuberculosis of the respiratory tract, but how should this be defined—sputum smear positivity or sputum culture negativity? If a compulsorily detained person with fully sensitive smear positive disease accepts standard short course chemotherapy,¹⁰ trial evidence shows that >90% should become smear and culture negative by two months and 98% culture negative by three months.11However, infectivity requiring segregation (if in hospital) is generally only required for two weeks because the infectivity of smear positive individuals declines rapidly.12 ,13 Therefore, even applying culture negativity, detention legally would be for a maximum of three months, only half the duration required for full treatment.10

The dilemma is even more complicated for HIV positive individuals or those with MDRTB. HIV positive individuals are much more susceptible to disease progression, perhaps 170 times that of HIV negative individuals,14 and in acquiring infection, so that even smear negative culture positive disease may be significantly infectious for this group. With MDRTB, because of the loss of the main killing drug (isoniazid) and the main sterilising drug (rifampicin), the usual rapid reduction in infectivity is no longer possible,12 ,13 and such individuals can remain infectious, however defined, for prolonged periods, sometimes lasting up to months.

The Government in its recent moves on Care in the Community for mental health announced alterations to the Mental Health Act to permit compliance orders which will force psychiatric patients to take their medication, and “assertive outreach teams” to police this with the right to compulsorarily readmit non-compliant patients. Whilst a person with smear positive tuberculosis not taking treatment, or taking it only intermittently, is not as immediately dangerous as an acute paranoid schizophrenic, such persons are infectious, transmit such infections readily to the unvaccinated and immunocompromised, if poorly compliant are at increased risk of developing and then transmitting drug resistance, tuberculosis still carries a significant morbidity and mortality even in immunocompetent cases (5859 cases in 1997, 392 deaths attributable to tuberculosis and 55 due to late effects; P Van Buynder, personal communication), and MDRTB carries a very much higher morbidity and mortality even in immunocompetent cases.15

A review of the powers for communicable disease control has been promised over the next few years when such issues will need to be debated by doctors and allied professions, patient representatives, lawyers and politicians representing the “public interest”. A possible pragmatic solution would be to increase the incentives to compliance, free drugs with practical help—food, housing, social support for disadvantaged groups such as the homeless and refugees (more carrot), but to strengthen or at least define clearly if and when compulsory detention (and treatment?) should be used for cases where the collaborative approach has failed (more stick). Such a system would be predicated on having minimum staffing levels to monitor and deliver treatment to recommended standards.10