The Performance of a Screening Test for Urgent Dental Treatment Need In

The performance of a screening IN BRIEF • Highlights the difﬁ culties in undertaking RESEARCH research in prisons. • Screening prisoners for dental conditions test for urgent dental treatment in ‘high turnover’ prisons is not effective. • Details the reasons why prisoners are a need in a prison population high need, difﬁ cult to engage population.

K. M. Buchanan,1 K. M. Milsom,2 L. Zoitopoulos,3 A. Pau4 and M. Tickle5

Objective To compare the performance of a questionnaire-based assessment of dental pain delivered by non-dental prison nursing staff against a clinical examination performed by an experienced prison dentist (gold standard). Set- ting The research fi eldwork took place in the healthcare department at HMP Brixton located in South London. Methods The cohort were those who had complained of dental/facial pain to the prison authorities and were therefore placed on the waiting list for emergency dental care. Subjects were asked to complete the screening test and were then clinically examined by trained and calibrated dentists. The screening test was in the form of a 12-item questionnaire designed to categorise a population reporting dental pain into one of three groups. The screening test results were compared with the diagnosis of a clinical examination. Results Ninety-six subjects were recruited during the 16-week study period. Of the 96 prisoners recruited, 27 of those failed to complete the screening test and/or the diagnostic examination even though they had reported pain previously . When sensitivity and specifi city values were calculated for the 96 prisoners recruited, the sensitivity was fairly high (81%) and the specifi city was poor (33%). However, when these values were calculated for the sub–population, those that completed both the screening test and diagnostic examination (n = 69), the sensitivity did not improve particularly (83%) but the specifi city value fell substantially to 13%. Conclusions This study highlighted the problems of conducting research in the prison environment, for example increased security preventing researchers from gaining access into the prison and general pressures on prison staff. Additionally, the study demonstrated that screening is not effective in local prisons with a high turnover of prisoners.

INTRODUCTION National Health Service (NHS). To sup- provide fast track access to dental serv- In 2003 the responsibility for funding port this process, in April 2006 primary ices for those screened positive. There prison health services in England was care trusts (local NHS bodies) were given are 24 and 48 hour general health checks transferred from the Home Offi ce to the responsibility for commissioning prison for newly admitted prisoners performed Department of Health. This was the fi rst health services and are now expected to by a nurse but there is no standardised step in a process of integrating prison ensure that prison health care is equiva- dental component to these assessments. health services into the mainstream lent to mainstream NHS services.1,2 A simple-to-complete questionnaire Many prison dental services are over- (Dental Pain Questionnaire (DePaQ)) stretched and this service is a signifi - has been developed and tested in the 1* 2 Research Manager, Consultant/Honorary Senior Lec- cant concern for PCTs and providers of NHS.7 This screening instrument clas- turer in Dental Public Health, Oral Health Unit, National Primary Care R&D Centre, University of Manchester, prison healthcare. Prisoners have high sifi es patients with dental pain into Higher Cambridge Street, Manchester, M15 6LP/Halton levels of untreated dental disease3-6 and three groups of common dental pain & St Helens Primary Care Trust; 3Consultant in Special Care Dentistry/Honorary Senior Lecturer in Special extensive oral neglect is common in conditions: Care Dentistry, Kings College London Dental Institute, newly admitted prisoners.2 Due to the Group 1 (acute periapical periodonti- Caldecott Road, London, SE5 9RW; 4Senior Lecturer, • Centre for Oral Adult Health, Barts and The London, extensive and unpredictable mobility tis and irreversible pulpitis) Queen Mary’s School of Medicine and Dentistry, Turner of remand and short stay prisoners and Group 2 (reversible pulpitis and den- Street, London, E1 2AD; 5Professor of Dental Public • Health and Primary Care, Oral Health Unit, National the restricted provision of dental serv- tine hypersensitivity) Primary Care R&D Centre, University of Manchester, ices in prisons, patients with toothache Group 3 (pericoronitis). Higher Cambridge Street, Manchester, M15 6LP • *Correspondence to: Miss Katie Buchanan can experience diffi culties in gaining Email: [email protected] rapid access to dental care.1 One possi- In a population attending an NHS Online article number E19 ble solution to this problem is to screen dental teaching hospital, the DePaQ was Refereed Paper - accepted 11 July 2008 new prisoners on entry to the prison capable of correctly classifying 89.7% of DOI: 10.1038/sj.bdj.2008.950 ©British Dental Journal 2008; 205: E19 for urgent dental treatment needs and dental pain cases initially categorised by

BRITISH DENTAL JOURNAL 1

clinical diagnoses. The respective sensitivity and speciﬁ city values of the ques- 96 prisoners reported having dental pain to the prison authorities, were subsequently tionnaire were Group 1: 0.80, 0.83, Group placed on the urgent dental treatment waiting list consented into the study period 2: 0.85, 0.89, and Group 3: 0.59, 0.90. If this tool can perform to a similar level in a prison population it has the potential 5 of the 96 subjects reported not having pain at the screening test to rapidly identify prisoners reporting pain who have an urgent treatment need without the need for all prisoners to see a dentist. This possibility could improve 7 of the remaining 91 subjects failed to complete the screening test questionnaire the lives of prisoners and help PCT and prison healthcare mangers plan dental service resources more efﬁ ciently. 15 of the remaining 84 subjects reported having no pain at the diagnostic examination stage The aim of this study was to assess the performance of the DePaQ in iden- tifying prisoners with an urgent need 69 subjects reported having dental pain and completed both the screening test for dental care when compared against and diagnostic examination, completed a questionnaire and a clinical examination clinical diagnoses.

METHODS Fig. 1 Flow diagram of prisoners through the study The study population was composed of male prisoners providing informed con- 3. Pericoronitis (inﬂ ammation con- in which two of the authors (LZ & AP) sent on the urgent treatment list of HMP nected to a partially erupted tooth, agreed the examination protocol and

Brixton in London. All of these prison- classically a wisdom tooth). diagnostic criteria for the common den- ers had complained of toothache to the tal pain conditions using case scenarios prison authorities previously and were Those subjects that had no pain, or in a classroom context. This was fol- listed to have rapid access to the dental the questionnaire was incomplete or lowed by a calibration exercise in which service in prison, which provided dedi- they did not complete a clinical exami- 25 patients were examined and diag- cated sessions to see prisoners complain- nation, were classifi ed into a separate nosed prior to the start of the study and ing of pain. The prisoners were invited group (group 4). Subjects categorised there was 100% agreement between the to participate by a trained member of into groups 1 (irreversible pulpitis, acute dentists. The outcome of each diagnostic the healthcare team when they arrived periapical periodontitis) and 3 (pericoro- examination was recorded on a simple at the dental department. HMP Brixton nitis) were adjudged to have an urgent clinical data collection form. The den- serves a number of courts in the London need of treatment, while those in the tists categorised the study population conurbation and houses a mixture of group 2 and group 4 were defi ned as into the same clinical groupings as the male remand and sentenced prisoners.8 having non-urgent treatment need. screening test. This prison houses up to 798 prisoners Due to the high levels of literacy prob- Analysis included simple frequency at any one time, with an a high turnover lems among prisoners,10 the question- distributions and cross tabulations. of prisoners, many staying just 35 days naire was read out to prisoners by the Responses to the questionnaire were on average.9 nurse who normally undertakes the 48 entered into classifi cation equations from Ethical approval was granted by South hour general assessment of new prison- a previous study using a dental hospital Essex Local Research Ethics Committee. ers. The language used to introduce the setting. Using these classifi cation equa- Prisoners are considered a vulnerable questionnaire and used in the question- tions, each prisoner was assigned to one group. The NHS ethical review process is naire itself was piloted with nursing of the three groups of clinical conditions in place to protect research participants staff and prisoners in HMP Brixton. and then recategorised into urgent or and the researchers. The conduct of the The prison nurse recorded the prisoner’s non-urgent groups. Cross-tabulations study ensured that prisoners were nei- response to each question on the ques- between questionnaire classifi cation ther coerced nor offered any incentive tionnaire, which took approximately and clinical diagnosis were carried out to participate. fi ve minutes to complete. to compute sensitivity and specifi city The screening test was in the form The diagnostic test consisted of a his- values for: of a 12-item questionnaire designed to tory and clinical examination defi ned • The whole population, that was all categorise a population reporting dental by a diagnostic protocol. Two experi- individuals consented into the study pain into three groups: enced prison dentists were trained in • A subpopulation, which comprised all 1. Irreversible pulpitis and acute peri- the diagnostic protocol and performed subjects who completed each stage of apical periodontitis (dental abscess) the diagnostic examinations blind to the the testing process, that is consent, 2. Reversible pulpitis and dentine outcomes of the questionnaire screening screening questionnaire and diagnos- hypersensitivity test. A training exercise was carried out tic examination.

2 BRITISH DENTAL JOURNAL

RESULTS Table 1 Frequency distribution of subjects at the screening test stage and diagnostic stage In total 96 subjects were recruited and consented. Figure 1 presents a fl ow WHOLE POPULATION n = 96 (%) diagram summarising the number of Screening questionnaire Diagnostic examination subjects at each stage of the fi eldwork. classifi cation classifi cation Five subjects admitted that they had no Irreversible pulpitis or acute periapical 70 (73) 44 (46) pain prior to administering the screen- periodontitis ing questionnaire, even though they Dentine hypersensitivity or reversible pulpitis 13 (14) 27 (28) had reported that they were suffering Pericoronitis 1 (1) 4 (4) toothache within the previous 24 hours. A further seven subjects could not com- Incomplete questionnaire 7 (7) - plete the questionnaire even though No pain reported 5 (5) 9 (9) a nurse was reading the questions to TOTALS 96 (100) 84 (100) them and fi lling in the questionnaire on their behalf. Following completion of the screening test, 15 subjects admit- Table 2 Sensitivity and specifi city for the whole population (n = 96) ted that they had not been experiencing pain once they had access to the exam- GOLD STANDARD Clinical diagnostic examination ining dentists. They then made requests Diagnosed URGENT Diagnosed NON-URGENT Totals for other forms of non-urgent dental (+) (-) treatment. This left 69 patients who had reported pain over the last 24 hours and Screened URGENT (+) 39 (55%) 32 (45%) 71 completed both a screening test and n (%) diagnostic examination. Table 1 shows the distribution of Screened NON-URGENT subjects screened positive and diag- 9 (36%) 16 (64%) 25 (-)

nosed positive according to the clini- Screening questionnaire n (%) cal categories. Irreversible pulpitis Totals 48 (50%) 48 (50%) 96 or periapical periodontitis were most commonly screened positive and diag- Sensitivity = 39 / (39 + 9) = 81% Specifi city = 16 / (16 + 32) = 33% nosed, whereas periocoronitis was rarely screened or diagnosed. There were large discrepancies between the outcomes of Table 3 Sensitivity and specifi city for the subpopulation who completed the whole the screening and diagnostic tests in the process, reported dental pain within 24 hours, competed screening test, completed diagnostic evaluation (n = 69) categories irreversible pulpitis or acute periapical periodontitis and dentine GOLD STANDARD Clinical diagnostic examination hypersensitivity or reversible pulpitis. Diagnosed URGENT Diagnosed NON-URGENT Totals These differences were evident in the (+) (-) sensitivity and specifi city values of the screening test. Screened URGENT (+) 39 (66%) 20 (34%) 59 Table 2 compares the outcome of the n (%) screening test with the outcomes of the diagnostic test for the whole population. Screened NON-URGENT Approximately three quarters of the 7 (70%) 3 (30%) 10 (-) population was screened positive. Sen- Screening questionnaire n (%) sitivity was relatively high (81%) but Totals 46 (66%) 23 (34%) 69 specifi city was poor (33%), with large numbers of subjects wrongly identi- Sensitivity = 39 / (39 + 7) = 85% Specifi city = 3 / (3 + 20) = 13% fi ed to the screening test as having an urgent treatment need. When the subpopulation that excluded individuals (Table 3), with large numbers of false research was initially scheduled to take who failed to complete a questionnaire positives being detected. place in two prisons, Brixton and Bel- or admitted to falsely reporting den- marsh. However, before the fi eldwork tal pain was examined, there was little DISCUSSION could begin, security in Belmarsh was improvement in performance of the test. The fi eldwork of this study illustrated signifi cantly tightened and fi eldwork in Sensitivity increased marginally (85%) some of the diffi culties in conducting this prison had to be abandoned as secu- but specifi city fell substantially to 13% research in a prison environment. The rity clearance for fi eldworkers could not

BRITISH DENTAL JOURNAL 3

be obtained. Staff at Brixton were con- proportion of those prisoners who were that the prisoners tended, in the ques- fi dent that they could recruit suffi cient screened as needing urgent treatment. tionnaire, to overestimate the severity subjects to complete the project within Therefore the instrument would have of their pain in order perhaps to obtain the planned timescales. We planned to some value if we are interested to ensure rapid care, rather than necessarily offer- recruit 200 subjects over a six week that those with urgent need are identifi ed ing untruthful answers. This inaccuracy period. It transpired that we recruited in the screening process. A low specifi - in pain self-reporting has been observed 96 subjects after four months. The fi eld- city means that those subjects identifi ed by others.14,15 Also, it could be that the work of the project was stopped as the as not needing urgent treatment were examining dentists underestimated the prison could not recruit the necessary classifi ed by the questionnaire as need- severity of the prisoners’ pain experi- adequately trained staff with the correct ing urgent treatment. The implication of ence or underestimated the prisoners’ security clearance to go into the prison this is that it may lead to inappropriate inability to cope and therefore assessed and escort prisoners to the healthcare scheduling of appointments for prison- them as not needing urgent treatment. facility for enrolment onto the project. ers needing dental care. Assessing the performance of a screen- Additional problems were encoun- Seven subjects were unable to com- ing test is complicated in this population tered, such as prison offi cers unilater- plete the questionnaire even though by diffi culties of making an accurate ally deciding to stop a research session the questions were read out to all par- diagnosis. Untreated dental disease is due to security concerns, and failure of ticipants and the responses recorded on almost ubiquitous in UK prison popu- prisoners to attend recruitment sessions. their behalf by a nurse. This may have lations16-18 and correct diagnosis of the This latter diffi cultly was surprising as been due to the failure to comply com- cause of dental pain requires the taking all prisoners scheduled for these vis- pletely with the process because subjects of an adequate history. The high preva- its had complained of toothache. Prison were not experiencing pain but wanted lence of untreated dental disease and staff reported that some prisoners were access to the dentist for other reasons. some subjects providing misleading his- reluctant to attend the healthcare facil- This problem was illustrated by the fact tories would hamper accurate diagnosis. ity due to fears about being attacked that 19 subjects said they had pain, but Provision of disingenuous answers to during transfers, and that other prison- then at the screening test (n = 4) or prior the screening test and uncertainty of ers failed to attend because they did not to the diagnostic examination (n = 15) diagnosis are likely to lead to haphaz- feel like attending the dental surgery admitted they were not experiencing ard outcomes, resulting in some false that day. However, the principal prob- pain but wanted access to the dentist to negative and some false positive results. lem hindering and ultimately prevent- ask for some other form of treatment, The low literacy levels10 may compound ing completion of the planned fi eldwork usually of a cosmetic nature, for exam- the haphazard nature of the outcomes was intermittent staffi ng problems in the ple a scale and polish or a crown. of screening and diagnostic tests in prison throughout the whole of the study When the 27 subjects who confessed this population. period. This problem was not due to lack to not having pain or could not complete Although the screening tool demon- of funding but due to the shortage of the questionnaire were excluded from strated excellent results in a population individuals with the necessary training the analysis, the specifi city remained attending an NHS dental teaching hos- and experience to supervise the trans- high but the sensitivity dropped to 13%. pital,7 it performed poorly in the prison fer of prisoners from the prison to the This could be attributed to subjects setting. Prior to undertaking this study, healthcare facility. There is no consensus ‘gaming’ the screening test, providing many experienced dentists working in in the literature on required sample sizes untruthful answers to again ensure that the prison service warned the research for psychometric validation of screen- they gained access to the dentist. For team that prisoners would be untruth- ing tests11-13 and although we fell short of example, some of the questions in the ful and claim that they were in pain just our target recruitment for the study, the screening test were related to duration to have time outside their cell or gain study provided interesting data. of pain. Many subjects reported that access to the dentist. This study quanti- The problems encountered during the they experienced pain over long peri- fi es this issue and strongly suggests that study period are problems that are often ods, perhaps to misleadingly convey screening is inappropriate for a local encountered daily in many healthcare the impression that they needed urgent prison with a high turnover population. departments across prisons in England. access to a dentist. This lay perception It should be remembered that screen- All of the issues mentioned above cause of the clinical presentation of dental ing is a process that includes screen- diffi culties in the day to day running of pain is at odds with the actual presen- ing, referral, diagnosis and provision of healthcare departments and ultimately tation of irreversible pulpitis or dental appropriate treatment, and the screen- hinder the smooth delivery of healthcare abscess. This pain is very severe but ing test is usually the easiest part of the services to this high need population. comparatively short lasting, therefore process to get right. Even if a screening Analysis of the whole population (n = the screening test would tend to cat- test produced high levels of sensitivity 96) showed that the test had good sensi- egorise subjects reporting that pain had and specifi city, screening could easily tivity but poor specifi city. Having high been present for a long period as having fail at other points in the process due to sensitivity means that the instrument non-urgent treatment needs. Another a highly mobile population and the part- was able to identify correctly a high interpretation of the fi ndings could be time nature of prison dental services.

4 BRITISH DENTAL JOURNAL

Screening may work in a less mobile, prison environment. There are a number health of remand prisoners in HMP Brixton, London. Br Dent J 2007; 202: E5. long-stay, convicted population of pris- of reasons for diffi culties encountered, 10. The Home Offi ce. Freedom of informa- oners. This study shows that the ques- for example, increased security prevent- tion release: literacy and numeracy levels of prisoners. 25 January 2005. Available at tionnaire is of little help in providing ing researchers from gaining access into http://www.homeoffi ce.gov.uk/about-us/free- effi cient access to care for newly admit- the prison and general everyday pres- dom-of-information/released-information/ foi-archive-offender-management/200-literacy- ted individuals with acute dental pain in sures on prison staff due to the shortage numeracy-prisoners?view=Html (accessed 8 high turnover local prisons. The National of staff. Additionally, the study demon- October 2008). 19 11. Hapak L, Gordon A, Locker D, Shandling M, Mock Screening Committee (NSC) acknowl- strated that screening is not effective D, Tenenbaum H C. Differentiation between mus- edges that screening is not the answer in local prisons with a high turnover culoligamentous, dentoalveolar, and neurologi- cally based craniofacial pain with a diagnostic to certain problems and that before a of prisoners. questionnaire. J Orofac Pain 1994; 8: 357-368. screening programme is introduced, ‘all 12. Galer B S, Jensen M P. Development and pre- 1. Department of Health, HM Prison Service. Strategy liminary validation of a pain measure specifi c to other options for managing the condition for modernising dental services for prisoners in neuropathic pain: the Neuropathic Pain Scale. should have been considered (eg improv- England. London: Department of Health, 2003. Neurology 1997; 48: 332-338. 2. Harvey S, Anderson B, Cantore S et al. 13. Thomas R J, McEwen J, Asbury A J. The Glasgow ing treatment, providing other services), Reforming prison dental services in England – Pain Questionnaire: a new generic measure of to ensure that no more cost effective a guide to good practice. Health Educ J 2005; pain; development and testing. Int J Epidemiol 64(Suppl. 4): ii-iii, 1-39. 1996; 25: 1060-1067. intervention could be introduced.’ For 3. Lunn H, Morris J, Jacob A, Grummitt C. The oral 14. Kikuchi H, Yoshiuchi K, Miyasaka N et al. Reliability this population of newly admitted short- health of a group of prison inmates. Dent Update of recalled self-report on headache intensity: 2003; 30: 135-138. investigation using ecological momentary assess- stay and remand prisoners, it may be more 4. Naidoo S, Yengopal V, Cohen B. A baseline survey: ment technique. Cephalagia 2006; 26: 1335-1343. appropriate to provide a clearly defi ned oral health status of prisoners – Western Cape. 15. Broderick J E, Schwartz J E, Vikingstad G, Prib- SADJ 2005; 60: 24-27. bernow M, Grossman S, Stone A A. The accuracy of symptomatic, pain relief-only service in 5. Osborn M, Butler T, Barnard P D. Oral health status pain and fatigue items across different reporting line with the guidance by the Department of prison inmates – New South Wales, Australia. periods. Pain 2008; 139: 146-157. 1 Aust Dent J 2003; 48: 34-38. 16. Bolin K, Jones D. Oral health needs of adolescents of Health. However, prison healthcare 6. Jones C M, Woods K, Neville J, Whittle J G. in a juvenile detention facility. J Adolesc Health managers must accept there will always Dental health of prisoners in the north west of 2006; 38: 755-757. England in 2000: literature review and dental 17. Young D S. Health status and service use among be ineffi ciencies in providing services to health survey results. Community Dent Health incarcerated women. Fam Community Health this diffi cult but high need population 2005; 22: 113-117. :1998; 21 16-31. 7. Pau A, Croucher R, Marcenes W, Leung T. Develop- 18. Walsh T, Tickle M, Buchanan K, Milsom K, and plan accordingly. It is nevertheless ment and validation of a dental pain-screening Zoitopoulos L. An investigation of the nature and important that we continue to work to questionnaire. Pain 2005; 119: 75-81. quality of research into dental health in prisons: A 8. HM Prison Service website. Brixton web- systematic review. Br Dent J 2008; 204: 683-689. fi nd a workable solution for this diffi cult page. http://www.hmprisonservice.gov. 19. UK National Screening Committee. Criteria to engage, high need population. uk/prisoninformation/locateaprison/prison. for appraising the viability, effectiveness and This study has highlighted the prob- asp?id=284,15,2,15,284,0 (accessed 8 appropriateness of a screening programme. October 2008). National Screening Committee, 2003. Available at lems of conducting research in the 9. Heidari E, Dickinson C, Wilson R, Fiske J. Oral http://www/nsc.nhs.uk/pdfs/criteria.pdf

BRITISH DENTAL JOURNAL 5