Patient and Health System Delays in the Diagnosis of Tuberculosis in Southern Thailand After Health Care Reform

Patient and health system delays in the diagnosis of tuberculosis in Southern Thailand after health care reform

M. Rojpibulstit,* J. Kanjanakiritamrong,† V. Chongsuvivatwong* * Epidemiology Unit, Faculty of Medicine, Prince of Songkla University (PSU), Hat-Yai, Songkhla, † Department of Pharmacy, Songkhla Hospital, Muang, Songkhla, Thailand

SUMMARY

SETTING: Seven public hospitals in a southern province was significant longer for patients with health insurance of Thailand. and first presenting to low-level public health facility OBJECTIVES: To measure delays in tuberculosis (TB) (i.e., community hospital, health centre, primary care diagnosis and to examine the factors associated with unit or private clinic/hospital). these delays, with special focus on the effect of drug CONCLUSIONS: The public should be informed how to store utilisation and health insurance coverage on pa- recognise TB symptoms to shorten patient delay. The tient delay. Thai National Tuberculosis Control Programme needs DESIGN: A total of 202 newly diagnosed smear-positive to supervise the private health sector, including drug and smear-negative pulmonary TB patients were inter- stores, for better TB control. Drug store personnel need viewed using a structured questionnaire. to be trained to recognise and refer TB suspects. The ca- RESULTS: The median patient, health system and total pacity of low-level public health facilities and private delay were 4.4, 2.8 and 9.4 weeks, respectively. Risk fac- doctors in TB diagnosis needs improvement. A proper tors for patient delay were age 31–60 years, having mild referral system should be developed. illness, previous similar symptoms and first presenting to KEY WORDS: tuberculosis; delay; drug store; health insur- non-qualified providers. Health insurance was not asso- ance coverage; Thailand ciated with a shorter patient delay. Health system delay

DELAYS IN TREATMENT of tuberculosis (TB) can METHODS result in a higher risk of mortality among patients and The protocol was approved by the Ethics Committees transmission of the disease in the community.1,2 Early of Prince of Songkla University and of the Ministry of detection and effective treatment, the two key factors Public Health, Thailand. in successful TB control, can be achieved by shorten- ing the time from the first symptom to arriving in standard health care (patient delay) and time between Study setting and health service systems first visit and diagnosis (health system delay). A first Songkhla province, the study population, was ap- visit to a drug store is common in developing coun- proximately 1.2 million. It is served by three tertiary tries,3–6 and has been reported to be a source of delay.3 hospitals, 15 community hospitals, 228 public health At the end of 2001, health system reform toward centres/primary care units (PCUs), 201 private clinics universal health care coverage (UC) was started in and six private hospitals. Of 270 drug stores in the Thailand.7 An understanding of whether reform has community, 94 operate without a pharmacist. shortened these delays would be an important lesson A TB clinic in each public hospital is supervised by for other developing countries where reform is also the National TB Control Programme (NTP). Accord- underway. ing to the referral system, UC patients must first at- The present study investigates factors associated tend a PCU or health centre to obtain free care at all with patient delay, with emphasis on the effect of levels of the service. Others, including those bypass- drug store utilisation and health insurance coverage. ing referral, receive free anti-tuberculosis drugs subsi- It also investigates health system delay, which might dised by the NTP, but must pay for medical investiga- be affected by the reform. tions and non-anti-tuberculosis drugs.

Correspondence to: Malee Rojpibulstit, Epidemilogy Unit, Faculty of Medicine, Prince of Songkla University, Hat-Yai, Songkhla, Thailand 90110. Tel: (66) 74-451165. Fax: (66) 74-212900. e-mail: [email protected] Article submitted 21 September 2005. Final version accepted 11 November 2005. Delays in diagnosis of TB, Southern Thailand 423

Study design and subjects ard ratio 1 suggests that the factor increases the risk Five community hospitals and two tertiary hospitals for delay. were selected; each had at least 10 new cases per 4- month cohort. All newly diagnosed pulmonary TB (PTB) patients RESULTS aged at least 15 years who were either sputum smear- Patient characteristics and site of first visit positive or smear-negative were included in the study. Of 212 initially recruited patients, 10 were excluded Those who were smear-negative and finally diag- due to uncertainty about the length of delay. The sub- nosed as non-TB, those diagnosed without sputum jects had a wide age range (15–83 years), and a mean examination and prisoners were excluded. age of 43.3 years (standard deviation 16.2). Over three quarters were male; 80% were Thai Buddhists. Data collection Two-thirds had finished only primary school or below, After obtaining written informed consent, a structured and the average monthly income (baht 5880, approx- questionnaire, pre-tested with 30 new TB patients imately US$150) was close to the national average and thoroughly revised, was used by a well-trained monthly per capita income (baht 5001, approximately researcher to interview each subject at the hospital US$125). Labourers with temporary employment shortly after diagnosis (up to 1 month). Five patients were most common (44%). Despite the declaration of declined to participate. Data were collected from June UC, more than a fifth of subjects were not covered by 2003 to April 2004. any insurance at the current visit, mainly due to their The questionnaire consisted of two parts. The first being a mobile population. part included socio-economic characteristics, health One-fifth of the patients were negative for AFB. At insurance status, clinical information on the current the time of the study, universal human immunodefi- illness and health service utilisation for the index symp- ciency virus (HIV) voluntary counselling and testing tom (cough). The second part, exclusively for drug for TB patients was not part of the national policy, store users, included details of drug store visits and and only 24% of subjects had benefited from this ser- drug store personnel practices. vice. Of 49 tested, 42.8% were positive. With such in- To minimise error in recalling patient delay, in ad- complete information, HIV status was not included in dition to structured sequencing of the time line, quali- the main analysis. tative probing questions were used. The interview data The drug store was the most common first site of were also checked against the data in medical records visit after the start of symptoms (43%); up to 12% for those patients who had previously visited a hospi- self-medicated from a grocery; only 31.5% sought tal with suspected TB symptoms. health care from public health facilities, and the private health sector occupied a minor share (13%). Definitions Adapted from previous studies,8,9 patient delay was de- Patient delay and determinants fined as the time from commencing cough to the first The median patient delay was 4.4 weeks (range 0.3– contact with a qualified provider (a doctor, health 33.1), with 61% reaching a qualified provider after centre or PCU). Health system delay was the time 3 weeks of cough. from first contact with the qualified provider to time The 86 initial drug store users had a median delay of diagnosis. Insured persons were patients receiving from first symptoms to drug store visit of 1 week. The free medical services, including government staff, pri- subsequent part of their patient delay had a median of vate employees under a social security scheme, people 6.4 weeks. After 3 weeks of cough, 52 made a repeat under the gold-card scheme (providing free care for visit to a drug store. elderly and children 12 years of age or co-payment From univariate analysis (Table 1), patients with of 30 baht [US$0.76] per visit for others) and those the following characteristics had a significantly longer holding government health cards and private health patient delay: male, age 31–60 years compared with insurance. younger age, AFB-positive, mild illness, being a smoker, The main independent variables were socio-demo- previous symptoms similar to the current illness, no graphic characteristics, health insurance status, acid-fast past experience of using the current visited hospital bacilli (AFB) status, severity of illness and place of and first presenting to non-qualified providers (drug first visit. store, grocery and traditional healer). Multivariate Cox regression analysis for patient Data analysis delay reduced significant predictors for prolonged The Mann-Whitney U and Kruskal-Wallis tests were delay to age 31–60 years compared with younger age, initially used to explore the association between delay mild illness, previous similar symptoms and first pre- and independent variables. Cox regression was then senting to non-qualified providers. Having health used to investigate factors associated with delay. A haz- insurance did not affect the time to seek help from 424 The International Journal of Tuberculosis and Lung Disease

Table 1 Univariate and multivariate analysis of factors associated with patient delay

Patient delay Median duration HR* Adjusted HR* Study variables n (IQR in weeks) (95%CI) (95%CI) Sex Male 156 4.5 (2.0–13.1) 1 1 Female 46 2.8 (1.1–7.0) 1.70 (1.21–2.39) 1.00 (0.68–1.46) Age group, years 15–30 53 3.3 (1.0–6.3) 1 1 31–45 73 5.0 (2.0–13.1) 0.58 (0.41–0.84) 0.64 (0.43–0.94)† 46–60 36 5.1 (2.7–15.9) 0.61 (0.40–0.94) 0.59 (0.37–0.93)† 60 40 3.4 (1.2–9.4) 0.76 (0.50–1.16) 0.86 (0.55–1.34) AFB status Positive 160 4.8 (2.0–13.0) 1 Negative 42 2.1 (1.0–4.4) 1.93 (1.36–2.73) Severity of illness‡ Mild 156 4.6 (2.0–12.9) 1 1 Moderate/severe 46 2.0 (1.0–4.8) 1.75 (1.25–2.44) 1.76 (1.22–2.55)§ Smoker Yes 128 4.8 (2.0–13.4) 1 No 74 3.1 (1.0–8.4) 1.48 (1.10–1.97) Previous similar symptoms Yes 39 5.3 (3.4–20.6) 1 1 No 163 4.3 (1.8–9.1) 1.49 (1.05–2.14) 1.52 (1.04–2.22)† Health insurance Yes 157 4.3 (2.0–11.0)1 1 No 45 4.6 (2.0–9.1) 1.06 (0.76–1.48) 1.41 (0.98–2.03) Site of first visit Qualified providers 91 2.0 (1.0–4.3) 1 1 Drug store 86 8.6 (4.0–21.6) 0.28 (0.20–0.39) 0.29 (0.20–0.42)§ Grocery/traditional healer 25 7.7 (4.4–11.7) 0.38 (0.24–0.60) 0.35 (0.22–0.56)§ Previous visit to current hospital Yes 146 4.3 (2.0–9.1) 1 No 56 5.1 (2.0–15.1) 0.71 (0.52–0.98)

* HR 1 suggests increasing risk of delay. † P 0.05. ‡ Classified into three levels according to whether the subject could do a full day’s work as usual (mild), only some daily activities (moderate) or nothing (severe). § P 0.01. IQR interquartile range; HR hazard ratio; CI confidence interval; AFB acid-fast bacilli. qualified providers on either univariate or multivari- pital or TB centre had a significantly shorter health ate analysis. system delay than those who first visited a lower-level health facility, i.e., community hospital, health centre, Health system delay and determinants PCU or private clinic/hospital. The health system delay was rather long, at a median On univariate analysis (Table 3), patients aged 46– of 2.8 weeks (range 0–74.4). After presenting to a 60 years had a significantly shorter system delay than qualified provider, TB suspects still made an average those aged 30 years. System delay at a tertiary hos- of 3.3 visits to the system before obtaining a final di- pital or TB centre was significantly shorter than at agnosis. Only 17 patients (8.4%) were diagnosed and lower-level health facilities or in the private sector treated at the first visit, 13 of them at a tertiary hos- (median 0.7 vs. 3.3–4.7 weeks). pital or TB centre. Slightly over one third (36.6%) of From multivariate Cox regression, the significant the patients were treated within 1 week after contact- predictors for shorter delay were aged 46–60 years ing a qualified provider. compared with age 30 years, having no health in- When stratified by type of site of first visit (Table surance and first presenting to a tertiary hospital or 2), health system delay contributed more than pa- TB centre. tient delay to the total delay in patients presenting initially to a qualified provider. However, these pa- Total delay tients had a significantly shorter total delay com- The median total delay was 9.4 weeks (range 0.8– pared to those who first presented to non-qualified 78.8). Only 16.3% of patients were treated within 4 providers. weeks after symptoms started. There was a weak but Of 91 patients first presenting to a qualified pro- significant negative correlation between patient and vider, patients whose first visit was to a tertiary hos- health system delay (Spearman R 0.22). Delays in diagnosis of TB, Southern Thailand 425

Table 2 Patient delay, system delay and total delay (weeks) by different sites of ﬁrst visit

Median Median Median n patient delay system delay total delay Overall (IQR) 202 4.4 (2–10.4) 2.8 (0.6–6.7) 9.4 (5.1–21.4) Sites of first visit Grocery/traditional healer 25 7.7 2.3 10.8 Drug store 86 8.6 2.1 13.1 Qualified providers 91 2.0 4.0 6.8 P value* 0.0001 0.0311 0.0001 Tertiary hospital†/TB centre 29 2.0 1.1 5.4 Community hospital 16 3.2 4.4 10.2 HC/PCU 19 2.0 4.7 6.8 Private clinic/hospital 27 1.0 5.0 7.6 P value‡ 0.1468 0.0217 0.2092

* Kruskal-Wallis test on median delays among three types of sites of first visit. † Includes general and regional public hospitals. ‡ Kruskal-Wallis test on median delays among patients first presenting to a qualified provider, stratified into four groups according to the level of health care. IQR interquartile range; TB tuberculosis; HC health centre; PCU primary care unit.

Drug store utilisation tor. After repeated visits, only 12 of 74 (16%) had Among 74 patients who could recall visit details, been recommended to see a doctor. none reported being given anti-tuberculosis drugs. Dispensed items included cough suppressants, muco- DISCUSSION lytics, bronchodilators and antibacterials. Thirty-four were asked at some time about their cough duration. This study shows substantial delays in the treatment Only four were advised at their ﬁrst visit to see a doc- of PTB patients in southern Thailand. Almost a quar-

Table 3 Univariate analysis and multivariate analysis of factors associated with health system delay

Health system delay Median duration HR* Adjusted HR* Study variables n (IQR in weeks) (95%CI) (95%CI) Sex Male 156 2.3 (0.4–5.9) 1 1 Female 46 4.1 (1.4–7.8) 0.79 (0.57–1.10) 0.86 (0.59–1.24) Age group, years 15–30 53 3.4 (1.0–6.8) 1 1 31–45 73 2.0 (0.4–6.0) 1.21 (0.85–1.72) 1.24 (0.85–1.80) 46–60 36 1.6 (0.4–3.6) 1.70 (1.11–2.62) 1.83 (1.18–2.84)† 60 40 5.1 (0.8–10.3) 0.78 (0.51–1.19) 0.80 (0.51–1.25) Ethnicity Buddhist Thai/Chinese 167 2.6 (0.4–6.3) 1 Muslim-Malaya 35 4.4 (1.0–12.3) 0.71 (0.49–1.03) AFB status Positive 160 2.8 (0.4–6.6) 1 1 Negative 42 2.5 (1.0–7.6) 0.91 (0.64–1.28) 1.13 (0.77–1.65) Smoker Yes 128 2.3 (0.4–5.6) 1 No 74 3.7 (1.0–9.3) 0.75 (0.56–1.00) Health insurance Yes 157 3.0 (0.7–6.8)1 1 No 45 2.3 (0.4–5.0) 1.31 (0.94–1.84) 1.52 (1.05–2.18)‡ First visit to qualified provider§ Tertiary hospital¶/TB centre 78 0.7 (0.3–3.3) 1 1 Community hospital 28 3.3 (0.5–10.7) 0.62 (0.40–0.96) 0.55 (0.35–0.86)† HC/PCU 21 4.7 (3.0–7.4) 0.57 (0.35–0.92) 0.56 (0.34–0.92)† Private clinic/ hospital 75 4.0 (1.7–6.8) 0.60 (0.43–0.82) 0.51 (0.36–0.72)† Patient delay (weeks) 202 2.8 (0.6–6.7) 1.11 (0.99–1.25) 1.07 (0.94–1.22)

*HR 1 suggests increasing risk of delay. † P 0.01. ‡ P 0.05. § Subjects were classified by first visit to qualified health facility, ignoring preceding visits to non-qualified facilities. ¶ Includes general and regional public hospitals. IQR interquartile range; HR hazard ratio; CI confidence interval; AFB acid-fast bacilli; HC health centre; PCU primary care unit. 426 The International Journal of Tuberculosis and Lung Disease ter of patients did not have health insurance; health appropriate anti-tuberculosis drug dispensing there- insurance was not associated with shorter patient fore seems less likely to occur in our setting. delay and it was independently associated with an In Thailand, free anti-tuberculosis drugs have been increased health system delay. Over half of patients provided for TB care over the past decade.25 How- initially relied on a drug store or grocery, which in- ever, free TB care cannot fully eradicate medical costs creased their delay. Presenting to a low-level official for patients. Under the non-UC system, large num- health facility added further delay to the health sys- bers of TB suspects who needed sputum examination tem delay. or chest X-ray had to bear the costs of such tests, and The median total delay of 9.4 weeks in this study is this may have been an obstacle to seeking health care.26 comparable to the reported range of 6.3–12.5 weeks If this is the case, UC should lead to better access to TB in Vietnam, Malaysia, Nepal, India, China and South care compared to NTP subsidies without UC. Africa (1997–2005).3,8,10–13 The reported mean total Having health insurance did not reduce the patient delays in previous Thai studies (1991–1997) range delay in our study. Over two fifths (41%) of those from 10 to 17 weeks compared to 15.8 weeks in our covered by health insurance still visited a drug store study.4,6,14 It is difficult to judge whether delays have first even though they had to pay. This finding is con- lessened over time, as the mean can be strongly af- sistent with recent publications on the impact of UC fected by outliers. on Thai people.27–29 Among a group of UC card hold- Without proper guidance and supervision, private ers who had acute illness in the past 2 weeks, 52% did health care facilities that are important first contact not use the benefit, including among their reasons the points for TB patients3,11–13 could be a cause of longer inconvenience in transportation, lengthy queues and treatment delay in some countries3,10–13 and provide lack of confidence in the quality of public health care. poor case management.3,10 In India and Vietnam, in- UC policy has been relatively successful in expand- volving the private sector in TB control resulted in an ing PCUs and upgrading the quality of health centre increase in case detection.15–17 services to improve accessibility and encourage people Drug stores in Thailand are an important facet of to utilise low-level public health facilities.30 Neverthe- the private health care sector. During 1997–2000, less, a first visit to a low-level health facility increased about one third (30–34%) of drugs consumed were the time to diagnosis by approximately 2.6–4.0 weeks distributed through drug stores.18 People use drug compared to a first visit to a high-level public health stores because of ease of access, availability of medi- care. Doctors are available at most health centres cines, no waiting times and convenient opening only once a week. Most PCUs are not equipped with hours.19 Currently, the ratio of drug stores with a TB diagnostic facilities, which has been reported as a pharmacist to the population is also rising.18 Drug factor associated with prolonged health system delay.31 stores will therefore continue to play an important Referral of TB suspects from these lower-level health role in the Thai health system. facilities to the hospital may be further delayed by the A first visit to a drug store increases the patient patients’ and their families’ unwillingness to go to the delay by approximately 6 weeks. A similar finding hospital. However, we had no data on severity of has been reported from Vietnam.3 Among 86 patients symptoms at the first point of contact with a qualified first presenting at a drug store, the greater the number provider. Patients with severe and more obvious symp- of drug store visits, the longer the delay before visit- toms may be more likely to present to a high-level ing a qualified provider. Sixty per cent of these pa- hospital or TB centre. tients still sought medications from a drug store after The UC-insured patients in our study had a longer 3 weeks of cough. In addition, only 5% (4/74) of drug system delay than the self-paying group. This may be store users reported that they were advised at their due to the fact that UC patients usually have to go drug store to see a doctor. Thus, delay associated with through the referral system, whereas self-paying pa- drug store visit is not explained solely by the early tients can obtain treatment promptly at their hospital symptoms when visiting. Patients’ and drug stores’ of choice. lack of awareness and suspicion of TB could probably As HIV status data were not complete, the effect of be the main cause. Financial incentives have been re- this variable on delays could not be determined. ported to play a part in irrational drug dispensing.20,21 Under the hospital-based setting, we could not inter- However, our study does not provide such evidence. view patients who were being treated in the private Further studies are needed to obtain clearer insight sector nor those who were not receiving any treatment into the mechanisms leading to the association of at all. However, these two groups are estimated to be patient delay and drug store use. small in Thailand.32 As in most studies on delay, data In South Asia, a substantial amount of anti-tuber- on length of delay can be subject to recall error. The culosis drugs were being sold in drug stores.22–24 Our data were, however, collected from patients who had study found that no TB patients had been dispensed just been diagnosed with TB under careful interview. anti-tuberculosis drugs. The development of multi- Errors were therefore minimised and were unlikely to drug-resistant Mycobacterium tuberculosis from in- have any differential bias that could distort the results. Delays in diagnosis of TB, Southern Thailand 427

CONCLUSIONS affecting delays in tuberculosis diagnosis in rural China: a case study in four counties in Shandong Province. Trop Med Int A comprehensive approach is needed to reduce treat- Health 2005; 99: 355–362. ment delay among TB patients. The public should be in- 14 Pungrassami P, Hirunyapa J, Tunsawai V, Pongpanich S, Petch- formed how to recognise TB symptoms to shorten pa- borisut O. The study of patient’s and doctor’s delay of tuber- tient delay. The NTP needs to supervise the private culosis case-detection in TB Center Zone 12 Yala. Thai J Tuberc Chest Dis 1993; 14: 73–83. health sector, including drug stores, for better TB con- 15 Arora V K, Sarin R, Lonnroth K. Feasibility and effectiveness trol. Training drug store personnel to recognise TB sus- of a public-private mix project for improved TB control in pects and developing a referral link from drug stores to Delhi, India. Int J Tuberc Lung Dis 2003; 7: 1131–1138. the NTP are essential. Improving the capacity of low- 16 Quy H T, Lan N T N, Lonnroth K, Buu T N, Dieu T T N, Hai level public health facilities and private doctors in TB di- L T. Public-private mix for improved TB control in Ho Chi Minh City, Vietnam: an assessment of its impact on case detec- agnosis as well as developing an effective referral system tion. Int J Tuberc Lung Dis 2003; 7: 464–471. for TB suspects from primary care centres is needed. 17 Kumar M K A, Dewan P K, Nair P K J, et al. Improved tuberculosis case detection through public-private partnership and Acknowledgements laboratory-based surveillance, Kannur District, Kerala, India, Financial support was provided by the Health Systems Research 2201–2002. Int J Tuberc Lung Dis 2005; 9: 870–876. Institute, the Royal Golden Jubilee PhD project and Prince of 18 Wibulpolprasert S, Gajeena A, Ekachampaka P, et al., eds. Songkla University. The authors would like to express their appre- Thailand Health Profile 1999–2000. Bangkok, Thailand: Bureau ciation to Dr P Pungrassami for her advice on the information of of Policy and Strategy, Ministry of Public Health, 2002. TB systems in Thailand, and Dr A F Geater who provided valuable 19 Wibulpolprasert S, Chokwiwat W, Thantiwate S, eds. Thai- comments on drafts of this paper. land Drug System. Bangkok, Thailand: Thailand Agricultural Cooperative Assembly Press, 2002. 20 Chuc N T, Larsson M, Falkenberg T, Do N T, Binh N T, Tomson References G B. Management of childhood acute respiratory infections at 1 Bustamante-Montes L P, Escobar-Mesa A, Borja-Aburto V H, private pharmacies in Vietnam. Ann Pharmacother 2001; 35: Gomez-Munoz A, Becerra-Posad F. Predictors of death from 1283–1288. pulmonary tuberculosis: the case of Veracruz, Mexico. Int J 21 Cederlof C, Tomson G. Private pharmacies and the health sec- Tuberc Lung Dis 2000; 4: 208–215. tor reform in developing countries-professional and commer- 2 Jochem K, Walley J. Determinants of the tuberculosis burden cial highlights. J Social Adm Pharm 1995; 3: 101–111. in population. In: Porter J D H, Grange J M, eds. Tuberculosis— 22 Hurtig A K, Pande S B, Baral S C, Porter J D, Bam D S. Anti- an interdisciplinary perspective. London, UK: Imperial Col- tuberculosis treatment in private pharmacies, Kathmandu lege, 1999: pp 33–48. Valley, Nepal. Int J Tuberc Lung Dis 2000; 4: 730–736. 3 Lonnroth K, Thuong L M, Linh P D, Diwan V K. Delay and 23 Lonnroth K, Lambregts K, Nhien D T, Quy H T, Diwan V K. discontinuity—a survey of TB patients’ search of a diagnosis in Private pharmacies and tuberculosis control: a survey of case a diversified health care system. Int J Tuberc Lung Dis 1999; 3: detection skills and reported anti-tuberculosis drug dispensing 992–1000. in private pharmacies in Ho Chi Minh City, Vietnam. Int J 4 Jittimanee S, Nasongkhla S, Vorasingha J, Wungmanee S. Tuberc Lung Dis 2000; 4: 1052–1059. Health seeking care of pulmonary tuberculosis patients in 24 Rajeswari R, Balasubramanian R, Bose M S, Sekar L, Rahman Thailand. Thai J Tuberc Chest Dis 1999; 20: 11–30. F. Private pharmacies in tuberculosis control—a neglected link. 5 Ngamvithayapong J, Yanai H, Winkvist A, Diwan V. Health Int J Tuberc Lung Dis 2002; 6: 171–173. seeking behaviour and diagnosis for pulmonary tuberculosis in 25 Payanandana V, Kladphuang B, Somsong W, et al. Battle against an HIV-epidemic mountainous area of Thailand. Int J Tuberc TB. National Tuberculosis Programme, Thailand, 1999. Bang- Lung Dis 2001; 5: 1013–1020. kok, Thailand: Tuberculosis Division, Department of Communi- 6 Tesana N, Chuin N, Supannawong P, Buphatrakul K. Delay of cable Disease Control, Ministry of Public Health, 1999. case finding in 6th Zonal Tuberculosis Center. Thai J Tuberc 26 Kamolratanakul P, Sawert H, Kongsin S, et al. Economic im- Chest Dis 1991; 12: 187–194. pact of tuberculosis at the household level. Int J Tuberc Lung 7 Universal Coverage Policy Committee. Universal Coverage. Dis 1999; 3: 596–602. Nonthaburi, Thailand: Health System Research Institute, 2001. 27 Chariyalertsak S, Roahasiriwong W, Suprakarnjanagunti S, et 8 Pronyk R M, Makhubele M B, Hargreaves J R, Tollman S M, al. A study of monitoring and evaluation of universal health Hausler H P. Assessing health seeking behaviour among tuber- coverage (UC) program phase I. Nuntaburi, Thailand: Health culosis patients in rural South Africa. Int J Tuberc Lung Dis Research System Institute, 2004. 2001; 5: 619–627. 28 Chariyalertsak S, Siwiroj P, Kaikwal J, et al. The evaluation of 9 Sherman L F, Fujiwara P I, Cook S V, Bazerman L B, Frieden circumstances for choosing a health service unit on voluntary T R. Patient and health care system delays in the diagnosis and basis in 5 pilot provinces. Nuntaburi, Thailand: Health Research treatment of tuberculosis. Int J Tuberc Lung Dis 1999; 3: System Institute, 2005. 1088–1095. 29 Naranong V, Naranong A. Expectation, health seeking behav- 10 Liam C K, Tang B G. Delay in the diagnosis and treatment of ior and impacts of Universal Health Coverage: people’s views. pulmonary tuberculosis in patients attending a university Nuntaburi, Thailand: Health Research System Institute, 2005. teaching hospital. Int J Tuberc Lung Dis 1997; 1: 326–332. 30 Health Care Reform Project. Situation of primary health care 11 Yamasaki-Nakagawa M, Ozasa K, Yamada N, et al. Gender dif- development in Thailand. Nuntaburi, Thailand: Ministry of ference in delays to diagnosis and health care seeking behaviour Public Health, 2002. in a rural area of Nepal. Int J Tuberc Lung Dis 2001; 5: 24–31. 31 Lawn S D, Afful B, Acheampong J W. Pulmonary tuberculosis: 12 Rajeswari R, Chandrasekaran V, Suhadev M, Sivasubrama- diagnostic delay in Ghanaian adults. Int J Tuberc Lung Dis niam S, Sudha G, Renu G. Factors associated with patient and 1998; 2: 635–640. health system delays in the diagnosis of tuberculosis in South 32 World Health Organization. Global tuberculosis control: sur- India. Int J Tuberc Lung Dis 2002; 6: 789–795. veillance, planning, financing. WHO report 2004. WHO/HTM/ 13 Cheng G, Tolhurst R, Li R Z, Meng Q Y, Tang S. Factors TB/2004.331. Geneva, Switzerland: WHO, 2004. 428 The International Journal of Tuberculosis and Lung Disease

RÉSUMÉ

CONTEXTE : Sept hôpitaux publics dans une province du nent pas d’un délai-patient plus court. Le délai du système Sud en Thaïlande. de santé est significativement plus long pour les patients OBJECTIFS : Mesurer les délais du diagnostic de la tuber- bénéficiant d’une assurance-santé et se présentant d’abord culose (TB) et examiner les facteurs associés avec ces à des services de santé officiels de bas niveau (hôpital délais, avec une attention particulière sur les effets de communautaire, polyclinique, unité de soins primaire ou l’utilisation de la pharmacie et ceux de la couverture par clinique/hôpital privé). les assurances-santé sur les délais-patient. CONCLUSIONS : Pour raccourcir le délai-patient, on SCHEMA : On a interviewé par un questionnaire struc- devrait former le public à reconnaître les symptômes de turé un total de 202 patients atteints de TB pulmonaire, suspicion de TB. Le Programme National de lutte contre à bacilloscopie tant positive que négative. la Tuberculose de Thaïlande doit superviser le secteur RESULTATS : Le délai médian du patient, du système de privé de santé y compris les pharmacies afin d’améliorer santé et le délai total ont été respectivement de 4,4 ; 2,8 la lutte antituberculeuse. Il est essentiel de former le per- et 9,4 semaines. Les facteurs de risque pour le délai- sonnel des pharmacies à reconnaître les cas suspects de patient sont un âge entre 31 et 60 ans, des symptômes TB et à les référer. Il est nécessaire d’améliorer la capa- peu importants, des antécédents de symptômes similaires cité des services de santé publics de bas niveau ainsi que et le fait de s’être présenté d’abord à des pourvoyeurs de celle des médecins privés en matière de diagnostic de la soins non qualifiés. Les assurances-santé ne s’accompag- TB. Un système de référence adéquate est nécessaire.

RESUMEN

MARCO DE REFERENCIA : Se estudiaron siete hospitales menor retraso por parte del paciente. El retraso del diag- públicos en una provincia del sur de Tailandia. nóstico dependiente del sistema de salud fue significati- OBJETIVOS : Medir el retraso en el diagnóstico de la vamente mayor en los pacientes que contaban con un se- tuberculosis (TB) y analizar los factores asociados con el guro médico y en quienes acudieron en primera instancia mismo, con especial interés en el efecto del recurso a las a un centro público de atención de salud de nivel básico farmacias y de la cobertura del seguro médico en el re- (hospital colectivo, policlínico, unidad de atencíon pri- traso dependiente del paciente. maria, clínico o hospital privado). MÉTODO : Se entrevistaron 202 pacientes con diagnós- CONCLUSIONES : Sería importante informar al público tico reciente de TB pulmonar, con baciloscopia posi- sobre los síntomas indicadores de TB, a fin de acortar el tiva o negativa del esputo, mediante un cuestionario retraso de los pacientes. Se precisa que el Programa Na- estructurado. cional de Lucha contra la Tuberculosis en Tailandia su- RESULTADOS : El retraso medio dependiente del pa- pervise el sector privado de salud, sin olvidar las farma- ciente fue 4,4 semanas, el retraso dependiente del sistema cias, con el objeto de optimizar el control de la TB. Un de salud 2,8 y el retraso global 9,4 semanas. Los factores aspecto primordial es educar al personal de las farmacias de riesgo para el retraso del paciente fueron la edad entre a fin de que reconozcan los pacientes con presunción di- 31 y 60 años, una enfermedad leve, el antecedente de agnóstica de TB y los remitan. Se requiere mejorar la ca- síntomas semejantes y el hecho de acudir en primera in- pacidad de las instituciones sanitarias de nivel básico y stancia a proveedores de atención de salud de baja cali- de los médicos del sector privado para el diagnóstico de ficación. El seguro médico no se correlacionó con un la TB ; se debe perfeccionar el sistema de referencia.