Open access, freely available online Correspondence Medical Decision Making:

The importance of a person-centred approach and the intricacies of risk communication have recently been well described in PLoS Medicine [1,2]. The applicability of the patient-centred approach to Eastern countries, however, has cultural, religious, and practical impediments that demand careful consideration. The bulk of the world population lives outside the United States and western Europe. Unlike in the West where the patient takes centre stage by both tradition and law, the family–doctor–patient triad is the norm in Eastern states, in general, and Pakistan in particular [3–8]. Pakistan is a predominantly Muslim country of 150 million people. About half the population is uneducated, and more than a third lives below the poverty line. There is one doctor for every 1,432 patients, compared to one doctor for every 390 patients in the US. The health-insurance system is virtually nonexistent, and there is no concept of assisted-care living, with the care of the elderly largely taking place at homes by their families. Strongly held religious beliefs and cultural views govern everyday life and dictate the roles of every member of the society. Families consist of well-knit, supportive, and collectively earning interdependent members who take mutual decisions on all matters pertaining to life and death [3,4]. The elder members of the family command the greatest respect and authority. The family unit is the functional unit of the society, the dynamics of which need attention and respect. Strong family systems and the authoritative position of the doctor are the governing forces of medical decision making in these countries. Illiteracy, poverty, poor awareness of patient's rights, and a lack of accountability for physicians are factors conducive to such a practice. With this background, the role of the patient is limited. Health expenditure is borne by the family, giving it a central role in decision making. The concept of the fi nancial survival of the family is a harsh reality [3,4]. The health-care costs of one seriously ill member may jeopardize the survival of others by draining the limited resources. Due to familial, moral, and monetary support, the patient relinquishes the responsibility of decision making and gives the primary role to the family or the doctor. Women, for example, may not give consent unless they get approval from their spouses [5]. In the case of women, who may have a lesser say in the patriarchal family system, the doctor should strive for active participation. Sometimes the family …

The importance of a person-centred approach and the intricacies of risk communication have recently been well described in PLoS Medicine [1,2].The applicability of the patient-centred approach to Eastern countries, however, has cultural, religious, and practical impediments that demand careful consideration.The bulk of the world population lives outside the United States and western Europe.Unlike in the West where the patient takes centre stage by both tradition and law, the family-doctor-patient triad is the norm in Eastern states, in general, and Pakistan in particular [3][4][5][6][7][8].
Pakistan is a predominantly Muslim country of 150 million people.About half the population is uneducated, and more than a third lives below the poverty line.There is one doctor for every 1,432 patients, compared to one doctor for every 390 patients in the US.The health-insurance system is virtually nonexistent, and there is no concept of assistedcare living, with the care of the elderly largely taking place at homes by their families.Strongly held religious beliefs and cultural views govern everyday life and dictate the roles of every member of the society.Families consist of well-knit, supportive, and collectively earning interdependent members who take mutual decisions on all matters pertaining to life and death [3,4].The elder members of the family command the greatest respect and authority.The family unit is the functional unit of the society, the dynamics of which need attention and respect.
Strong family systems and the authoritative position of the doctor are the governing forces of medical decision making in these countries.Illiteracy, poverty, poor awareness of patient's rights, and a lack of accountability for physicians are factors conducive to such a practice.With this background, the role of the patient is limited.Health expenditure is borne by the family, giving it a central role in decision making.The concept of the fi nancial survival of the family is a harsh reality [3,4].The health-care costs of one seriously ill member may jeopardize the survival of others by draining the limited resources.Due to familial, moral, and monetary support, the patient relinquishes the responsibility of decision making and gives the primary role to the family or the doctor.Women, for example, may not give consent unless they get approval from their spouses [5].In the case of women, who may have a lesser say in the patriarchal family system, the doctor should strive for active participation.Sometimes the family aims to protect the patient from stress by withholding information, and in terminal illnesses, the doctor and family act in concert to conceal complete information from the patient.They, for example, may not mention the word cancer to patients who have malignancies [9].
In contrast to Western practice, the role of the doctor is authoritative.Doctors are regarded as instruments of God and given the fi nal authority in decision making [3,4,10].In such circumstances, doctors are likely to take decisions unilaterally.When they do involve patients in decision making, physicians accept the centrality of families, with some considering patients and families as one [5].One worry regarding communication of harm is of losing patients to other physicians with a more reassuring "nothing will go wrong" attitude [5].It is also said that more time and patience are required to explain things to the illiterate.It is perhaps impractical, therefore, to expect overworked and underpaid physicians to practice risk communication according to the book.
Thus, the concept of individual centrality that is so elementary in the West stands challenged in the East.Research is needed to formulate appropriate strategies of risk communication.Areas needing research include the patient's concept of autonomy, the role of the family as perceived by patients and doctors, the existing practices of medical decision making, and the training of doctors in communicating risk.
An economically sound and literate population, properly trained doctors, and institutional checks and balances are essential prerequisites for establishing decision making with parity of partners.The need is to fi nd a middle ground where not only the family unit is respected, but the patient also plays a proactive role.A dynamic balance between cultural values of caring and the possibility of a more individualistic role in health care is needed and is, indeed, attainable.Doctors, being the most infl uential component of the family-doctorpatient triad, can play a signifi cant role in bringing about this change.

Malaria Diagnosis and Treatment: One Size Does Not Fit All Chris Drakeley, Roly Gosling, Hugh Reyburn
Icke and colleagues highlight an online and CD-ROM source available for the teaching of malaria diagnosis [1].While this is a commendable initiative that provides open access to badly needed training materials, we feel that it is overambitious to propose this Web site (which gives equal weight to diagnosis, treatment, and chemoprophylaxis) as a resource that can be equally relevant to the issues of travel-related malaria, to which the Web site primarily applies, and the problems of malaria diagnosis and treatment in sub-Saharan Africa, which are described as the primary focus in the PLoS Medicine article [1].
The microscopic diagnosis of malaria is indeed an important resource in Africa to which the large majority of health facilities do not have access.Even where microscopy is available, there is evidence of substantial overdiagnosis of malaria and that better access to malaria microscopy may not result in better targeting of antimalarials [2,3].What data there is on the accuracy of slide results suggest that microscope quality, preparation of blood fi lms, and quality of reagents are at least as important in constraining the quality of results as the ability of slide readers to identify Plasmodium parasites from high-quality thin blood fi lms [4][5][6].In fact, thick blood fi lms are the norm in Africa, thin fi lms being almost exclusively confi ned to research, yet these are in the small minority on the Web site.
P. falciparum is overwhelmingly the dominant Plasmodium species seen in Africa, and training guides that try to give a "global overview" run the risk of giving a misleading impression of the relative importance of the four species in any given setting where malaria is endemic.The only actual setting where all species of malaria might be seen relatively frequently would be in travel clinics in developed countries.This point might seem obvious, but for many junior laboratory technicians in Africa hungry for scarce training opportunities it is potentially distracting and confusing.
The problems of simultaneously addressing the issues of malaria in endemic areas of Africa and in travellers from resource-rich countries are particularly applicable to malaria treatment.On the Web site, among the drugs recommended for the treatment of non-severe malaria are quinine, atovaquone/proguanil hydrochloride, mefl oquine, and two tetracyclines.In Africa quinine is reserved for severe malaria and is associated with a high rate of failure due to poor adherence in outpatients; atovaquone/proguanil hydrochloride and mefl oquine are not widely used because of their high cost; and tetracyclines are contraindicated in children and pregnant women, who bear the overwhelming burden of malaria.
The Web site describes criteria for severe malaria that are not appropriate for Africa.Thus, cerebral malaria, jaundice, renal failure, and lactic acidosis (impossible to detect in most African hospitals where "respiratory distress" is the equivalent criterion) are all listed ahead of severe anaemia, the most common manifestation of severe malaria in Africa.Repeated convulsions are not mentioned, presumably because they are rare among travellers with malaria, but they are common among children in Africa and are associated with a poor outcome.Renal failure as a manifestation of severe malaria is mentioned second but is very uncommon in Africa [7,8].
The apparently large numbers of African health-care workers who have accessed this site might legitimately feel confused by some of these descriptions, and there is no indication of which, if any, sections apply to Africa and which to travellers.While many will interpret the information critically, others may not, especially when English is not their fi rst language and where there is a tendency to accept didactic sources in preference to what is obvious from personal or local experience.
The diagnostic component of the Web site has important potential as a training tool, but we feel that signifi cant modifi cation and more explicit indication of information that is country-specifi c are needed before its use is likely to result in improved standards of care in African hospitals and clinics. It

Authors' Reply: Teaching Health Workers Malaria Diagnosis
With reference to the letter commenting on our paper [1] from Drakeley and colleagues [2], they have some justifi cation in stating that one size does not fi t all.However our Web site, which is the basis for their comment, was originally designed on a wholly voluntary basis for Australia.We have been overwhelmed by the interest and acceptance of this site worldwide.Not only have we had more than 750,000 visitors to the site but have issued free CD-ROMs to institutions in 149 countries.
The fact that it was warmly embraced by so many others from around the world bears testimony to its usefulness as, indeed, do the tens of thousands of letters and Emails thanking us for our efforts.We have made some modifi cations on our annual update in response to suggestions and changes in approach.The main section of interest has been the section on diagnosis, testing, and teaching, and perhaps the reason for this is the high quality of the illustrations.One only has to look at the site to recognize that we have not given equal weight to all sections, as Drakeley and colleagues point out.The emphasis is on diagnosis, testing, and teaching.Many large organizations have requested a substantial number of copies of the CD-ROM, and in Germany one organization has been printing its own (with permission).We are aware of some superb CD-ROMs on malaria put out from other sources, but they are expensive for organizations with a very small budget.
We accept that diagnosis by thick fi lm is the norm in Africa and a number of other countries and, in fact, have spent many years ourselves diagnosing malaria from thick fi lms in India and Southeast Asia.We have described how to make thick fi lms and have provided a picture.We have mentioned the staining of fi lms, but we have not described how to prepare the stains because we considered that outside our brief.It is important with Web sites to be concise, otherwise they won't be read.The actual diagnosis of malaria is the same for Africa, India, South America, and Southeast Asia, and it is the proper diagnosis that we believe is paramount.We know that language can be a serious problem.We have provided a version in French and Spanish.The French version we are told is useful for certain parts of Africa.When other languages have been requested, we have suggested that a small booklet should be written in the local language by those with local knowledge.
In regards to the comments on treatment, this section was written by T. M. E. Davis, who holds the Chair of Medicine at the University of Western Australia and is a consultant on malaria to Thailand and Cambodia.If we sought an expert on treatment for every endemic region, we would never get the material into print.One is not always able to use the drugs of choice in Southeast Asia because up to 50% of antimalarials sold in some areas are fake.It has been stated that children should not be given tetracyclines, but that has already been made very clear on the site.
We are aware that the site needs to take into account various interests and situations, which is why we have included our E-mail addresses on the site.We presume that as experts in the fi eld Drakeley and colleagues would have been aware of the site either in Tanzania or the London School of Tropical Medicine, where a number of CD-ROMs have been requested and sent.We hope and expect that eventually a group with the enthusiasm of these correspondents will accept the challenge and produce a site that will overcome the problems that are the cause of their concern.We fully understand the diffi culty of doing this on a voluntary basis.In the meantime, we will continue to service the site and hope that the very large number of users will continue to fi nd it helpful.Finally, we state once again that if concise suggestions for improvements are sent to us by E-mail we will give them serious consideration.Our E-mail contact is now sandy.treadgold@health.wa.gov.au.

Richard E. Davis Royal Perth Hospital Perth, Western Australia, Australia
Recently PLoS Medicine published our paper entitled "Designing Equitable Antiretroviral Allocation Strategies in Resource-Constrained Countries" [1].We were disappointed to fi nd that the editorial perspective written by the World Health Organization (WHO) ethicists regarding our paper [2] was based upon a substantial misunderstanding of our novel quantitative analyses and our important results.Hence, they misunderstood the signifi cance of the health-policy implications of our results.Thus, we wish to correct the record.
Firstly, Capron and Reis [2] misunderstood our quantitative analyses.They stated that "Wilson and Blower developed a mathematical model that could inform policymakers' decisions regarding the optimal distribution of treatment sites to ensure equal access by all individuals infected with HIV."However, our model does not determine the optimal distribution of treatment sites.As we clearly state in our paper [1] (and is also stated in the synopsis [3]), we developed a model that policy makers can use to make decisions regarding how to achieve the optimal allocation of scarce antiretrovirals among the available health-care facilities (HCFs) if the objective is to ensure treatment equity.We also calculated how the optimal allocation of antiretrovirals would vary if the number of HCFs utilized increased and/or the size of the catchment area that each HCF services increased [1].Thus, we took the treatment sites (i.e., HCFs) as given, and we used their specifi c spatial location in South Africa as inputs to our model in order to determine optimal antiretroviral allocation strategies under a variety of conditions.
Secondly, Capron and Reis [2] misunderstood our important results.They stated that "applying this tool to the South-African province of KwaZulu-Natal, Wilson and Blower were able to confi rm mathematically the intuitive assumption that using a maximum number of centers, at the least possible distance from most affected populations, would lead to the greatest fairness in the geographical distribution of ART [antiretroviral therapy]."We agree that if these had been our results, they would have been trivial and obvious.However, Capron and Reis [2] did not discuss our actual results: we determined how to decide how many drugs to allocate to each of the available HCFs in order to achieve an optimal allocation if the objective is to ensure treatment equity.This is a very complex problem and the antiretroviral allocation strategies that we calculated (by using our model) to be optimal are very complex (see Figure 3 in our paper, which graphically shows the proportion of drugs that should be allocated to each of the available HCFs).Furthermore, we also determined what catchment area each HCF should service; specifi cally, we calculated that each HCF should serve (if the objective is to achieve treatment equity) a catchment area of 40-60 km.Thus, our results demonstrate (to our knowledge for the fi rst time) that patients infected with HIV will have to travel extremely large distances (i.e., 40 -60 km) in order to receive antiretrovirals, if the objective is to achieve treatment equity in South Africa.We stress that currently it is unknown what the actual size of the catchment area is around HCFs in South Africa.Catchment areas may in fact be very small.Thus, we suggested [1] that a primary goal should be to obtain empirical data of the distances that patients in South Africa are willing (or able) to travel in order to receive antiretrovirals.We have been the fi rst to provide a quantitative assessment of the necessary size of the catchment area, and our results have identifi ed that there is an urgent need to collect these critical data for quantifying the size of the catchment areas around HCFs.We have determined that the size of the catchment area will be a critical component in the ability to achieve treatment equity in South Africa.We also compared the optimal antiretroviral allocation strategies that we calculated with the current plan of the South African government for allocating antiretrovirals [4], and we determined that the current antiretroviral allocation strategies in South Africa will not achieve treatment equity.Taken together, our quantitative results are novel and controversial, providing important quantitative insights into a complex public-health problem.
We applaud the ambitious "3 by 5" WHO target for the antiretroviral rollout.However, the WHO has not yet devised a quantitative policy for determining how to allocate antiretrovirals in situations where the demand for drugs greatly exceeds the supply [5].Health-policy offi cials in each country will have to make these important and diffi cult decisions, and they will all make different decisions based upon what objectives they wish to optimize and prioritize.There are a multitude of factors to consider (these factors are well described in the recent Institute of Medicine report [6]).We stress that the alternative to a quantitative rational approach for allocating scarce resources is an ad hoc approach, which is how the scarce supply of antiretrovirals is currently being distributed in many resource-constrained countries.Our operations research modeling approach is based upon spatial heterogeneity in the distribution of HCFs in South Africa and the spatial heterogeneity of the HIVinfected population.The most important "real world" result is that we show that what the South African government is currently doing is inequitable.We show them how to achieve equity, if they wish to do so.We hope that our novel approach for deciding how to allocate antiretrovirals will be of use to the WHO and also to the relevant authorities in the many resource-constrained countries who will soon have to make very diffi cult decisions as to who lives and who dies.Our analysis is to our knowledge the fi rst analysis to show how a rational and scientifi c solution can be reached for deciding how to allocate a limited amount of antiretrovirals, if the goal is to achieve treatment equity.Clearly, other goals must be taken into consideration (and our model can be modifi ed to include these other goals); however, we hope that treatment equity will be a very high priority during the antiretroviral rollout that is just beginning.

David P. Wilson
Sally M. Blower (sblower@mednet.ucla.edu)University of California Los Angeles, California, United States of America On a factual level, we believe Wilson and Blower's results were appropriately described for the purposes of our commentary.As their letter points out, they applied the operations research methods to model the allocation of antiretrovirals (ARVs) among 17 health-care centers in KwaZulu-Natal, based on a hypothetical distribution of HIV/AIDS among the communities in that province.Their article characterizes this as "an elegant and simple theoretical framework," but they object to our concluding that it could "inform policy-makers' decisions regarding the location of HIV services," since they took the treatment sites as given.Yet their article compared the alternatives of using all 54 centers in the province, at one extreme, and of using only a single treatment site (in Durban), at the other extreme; in each case, the possibility of allocating to a larger number of centers is equivalent to the creation of additional centers closer to remote groups of patients.
Wilson and Blower write that geographic accessibility is improved if the number of health-care facilities is increased, and they calculated that it would be optimal if all 54 facilities in the province of KwaZulu-Natal distributed the medicines, instead of just 17.We took this result to confi rm the need to reach out and build capacity.We are sorry if we were mistaken in assuming that Wilson and Blower would want to see their stated objective of ensuring fair distribution applied in the realworld context of many poor countries with a high HIV burden and where fairness in ARV care cannot be achieved solely by allocating resources among the existing sites.
A wider gap in perception can be seen in Wilson and Blower's repeated confl ation of "optimal," "equal," and "equitable," combined with their suggestion that decision makers who fail to apply their model must be following an "ad hoc approach."The central point of our commentary was that various ethical theories reach very different conclusions about what result would be optimal, and that even among those aiming to achieve the greatest equity (rather than some other optimum), many would not take equality as the measure of equity.Wilson and Blower themselves recognize that apparent equality of access (in terms of distance to treatment) needs further study to determine whether patients can in fact access treatment.We need to know whether some distances are simply too far for patients to travel for chronic care, and when distances of equal length affect access very differently because of the characteristics of particular patient populations, transportation systems, and so forth.
Wilson and Blower seem unwilling to accept the notion that, in the furtherance of a rational strategy to achieve equity, some health authorities might decide, for example, to allocate a disproportionate share of ARVs to traditionally disadvantaged populations.Wilson and Blower's model could still be useful in allocating resources among the centers chosen (or established) to reach the target population, but the calculation would have to take account of more information about the centers and the population, lest assumptions about catchment areas produce a formal equality that does not translate into actual equality in access, much less into equitable access in light of all relevant factors.
Plainly, we share Wilson and Blower's aim of optimizing countries' responses to the tragedy of treatable, but untreated, HIV/AIDS.Any tools that are useful to that end are welcome.But besides using models to distribute ARVs in a way that optimizes spatial equality, governments that want to achieve equity will need also to overcome nongeographic barriers to accessing treatment.These include ignorance, stigma, discrimination, and outright criminalization of vulnerable groups, as well as fees at point of service that are prohibitive for the poor.All of these are given attention within the context of the "3 by 5" program of the World Health Organization and the United Nations Joint Programme on HIV/AIDS, including in the guidance document on equitable access to ARV treatment cited in our commentary [4].

The Debate over Placebo-Controlled Trials Frankllin Miller
Turner and Tramèr provide a cogent argument in favor of the ethical use of placebo controls despite "proven effective treatment" [1].However, they are wide of the mark citing the APPROVe trial in support of their position.Because there is no established treatment to prevent adenomatous polyps, few commentators would have any objections to the use of placebo controls in this study.Nevertheless, they are right to suggest that it would have been desirable to have included a placebo control in the VIGOR study to provide a more rigorous assessment of safety.Whether, all things considered, a placebo control would have been ethical in this study of treatment for rheumatoid arthritis is debatable.
Another issue not discussed in this PLoS Medicine Debate is the value of placebo controls in early "proof of concept" effi cacy trials, despite the existence of established treatment.The effi ciency of seeking a rigorous effi cacy signal before moving on to larger-scale trials (and exposing as few subjects as possible to drugs that might not work or turn out to be toxic) is a valid ethical reason for using placebo controls, provided subjects are not exposed to undue risks of harm from withholding established treatment [2].
Capron AM, Reis A (2005) Authors' reply.PLoS Med 2(6): e186.Copyright: © 2005 Capron and Reis.This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Competing Interests: The authors have declared that no competing interests exist.DOI: 10.1371/journal.pmed.0020186 is true that Internet technologies can "revolutionise information technology…for healthcare professionals in developing countries".The World Health Organization and others already have very extensive Web sites that are contributing to this.Searching the Internet on "malaria diagnosis" reveals a number of sites, many with excellent sections, but one is struck by the lack of clarity defi ning the target audience and context of the problem.For much information available on the Web this may not matter, but for malaria we feel it does.